Capgemini Data Science Project - Recommendation System

Introduction:

This notebook covers the Hackathon problem.
Problem Statement: Develop a recommendation model from the given 2 datasets.
Datasets given: Electronics Sales & Modcloth Sales.

For Recommendation, I have made 3 criteria:
a. Recommendation based on Ratings
b. Recommendation based on Demand of Items/Number of Units Sold
c. Recommendation based on Market Basket Analysis (using Apriori)

Procedure:

1. Importing Libraries & Preliminary Data Exploration
   
   
2. Data Cleaning
   
   
3. Feature Engineering & Exploratory Data Analysis
   
   
4. Recommending Items based on Ratings
   a. Recommending Items
   b. Recommending Brandwise Items
   c. Latest Trend: Recommending Items over past 2 years
   d. Latest Trend: Recommending Brandwise Items over past 2 years
   
   
5. Recommending Items based on Demand of Items/Number of Units Sold (i.e. Best Selling items)
   a. Recommending Items
   b. Recommending Brandwise Items
   c. Latest Trend: Recommending Items over past 2 years
   d. Latest Trend: Recommending Brandwise Items over past 2 years
   
   
6. Applying Market Basket Analysis (using Apriori algorithm)
   a. Recommending items based on items purchased in past
   b. Recommending categories based on item category purchased in past
   
   
7. Conclusion

1. Importing Libraries & Preliminary Data Exploration

Libraries:

We will be making use of below libraries:

* Pandas: for reading & manipulating data via forming DataFrame <br>
* Plotly: for data visualisation <br>
* Mlxtend: for making association rules (market basket analysis) between items and categories

"""installing libraries"""
# !pip install plotly
# !pip install mlxtend

'installing libraries'

import pandas as pd
import plotly.express as px
from mlxtend.frequent_patterns import apriori, association_rules
from mlxtend.preprocessing import TransactionEncoder
import mlxtend
import numpy as np
from scipy.sparse import csr_matrix
from sklearn.neighbors import NearestNeighbors

Preliminary Data Exploration

"""Reading the Datasets: We form dataframes from the 2 .csv files"""
raw_electronics_db = pd.read_csv("electronics.csv")
raw_modcloth_db = pd.read_csv("modcloth.csv")

"""Dataframe of Electronics database"""
raw_electronics_db.head()

	item_id	user_id	rating	timestamp	model_attr	category	brand	year	user_attr	split
0	0	0	5.0	1999-06-13	Female	Portable Audio & Video	NaN	1999	NaN	0
1	0	1	5.0	1999-06-14	Female	Portable Audio & Video	NaN	1999	NaN	0
2	0	2	3.0	1999-06-17	Female	Portable Audio & Video	NaN	1999	NaN	0
3	0	3	1.0	1999-07-01	Female	Portable Audio & Video	NaN	1999	NaN	0
4	0	4	2.0	1999-07-06	Female	Portable Audio & Video	NaN	1999	NaN	0

"""Dataframe of Modcloth database"""
raw_modcloth_db.head()

	item_id	user_id	rating	timestamp	size	fit	user_attr	model_attr	category	brand	year	split
0	7443	Alex	4	2010-01-21 08:00:00+00:00	NaN	NaN	Small	Small	Dresses	NaN	2012	0
1	7443	carolyn.agan	3	2010-01-27 08:00:00+00:00	NaN	NaN	NaN	Small	Dresses	NaN	2012	0
2	7443	Robyn	4	2010-01-29 08:00:00+00:00	NaN	NaN	Small	Small	Dresses	NaN	2012	0
3	7443	De	4	2010-02-13 08:00:00+00:00	NaN	NaN	NaN	Small	Dresses	NaN	2012	0
4	7443	tasha	4	2010-02-18 08:00:00+00:00	NaN	NaN	Small	Small	Dresses	NaN	2012	0

---

Checking the datasets by number of purchases made by every user

Users in Electronics Database

"""Number of purchases made by every user"""
raw_electronics_db.user_id.value_counts()

142967    41
30661     38
89185     37
80476     34
46878     34
          ..
702412     1
692171     1
696265     1
694216     1
0          1
Name: user_id, Length: 1157633, dtype: int64

from the above, we see that there are in all 1157633 users in electronics database

---

"""cases where the user has purchased only 1 item"""
raw_electronics_db.user_id.value_counts()[
    raw_electronics_db.user_id.value_counts() == 1
]

533596    1
944283    1
374253    1
252184    1
196866    1
         ..
702412    1
692171    1
696265    1
694216    1
0         1
Name: user_id, Length: 1054823, dtype: int64

From the above, we see that there are around 1054823 cases where the user has purchased only 1 item.
For remaining users, we can form a market basket analysis modelling

---

Users in Modcloth Database

"""total number of users in Modcloth database"""
raw_modcloth_db.user_id.value_counts()

Sarah            250
Amanda           204
jennifer         198
Laura            197
Jennifer         191
                ... 
contact            1
lisamarie1019      1
auraroy            1
kaylabr            1
lenab              1
Name: user_id, Length: 44783, dtype: int64

from the above, we see that there are in all 44783 users in modcloth database

---

"""cases where the user has purchased only 1 item"""
raw_modcloth_db.user_id.value_counts()[raw_modcloth_db.user_id.value_counts() == 1]

ram.z.rach           1
kyliemwilliams074    1
emma.c.newhall       1
ashlatash2297        1
Manager              1
                    ..
contact              1
lisamarie1019        1
auraroy              1
kaylabr              1
lenab                1
Name: user_id, Length: 30307, dtype: int64

From the above, we see that there are around 30307 cases where the user has purchased only 1 item.
For remaining users, we can form a market basket analysis modelling

---

2. Data Cleaning

Let us check for missing values for Electronics dataframe

"""For electronics Database"""
raw_electronics_db

	item_id	user_id	rating	timestamp	model_attr	category	brand	year	user_attr	split
0	0	0	5.0	1999-06-13	Female	Portable Audio & Video	NaN	1999	NaN	0
1	0	1	5.0	1999-06-14	Female	Portable Audio & Video	NaN	1999	NaN	0
2	0	2	3.0	1999-06-17	Female	Portable Audio & Video	NaN	1999	NaN	0
3	0	3	1.0	1999-07-01	Female	Portable Audio & Video	NaN	1999	NaN	0
4	0	4	2.0	1999-07-06	Female	Portable Audio & Video	NaN	1999	NaN	0
...	...	...	...	...	...	...	...	...	...	...
1292949	9478	1157628	1.0	2018-09-26	Female	Headphones	Etre Jeune	2017	NaN	0
1292950	9435	1157629	5.0	2018-09-26	Female	Computers & Accessories	NaN	2017	NaN	0
1292951	9305	1157630	3.0	2018-09-26	Female	Computers & Accessories	NaN	2016	NaN	0
1292952	9303	1157631	5.0	2018-09-29	Male	Headphones	NaN	2018	NaN	0
1292953	9478	1157632	1.0	2018-10-01	Female	Headphones	Etre Jeune	2017	Female	0

1292954 rows × 10 columns

"""missing values per column are as below"""
raw_electronics_db.isna().sum()

item_id             0
user_id             0
rating              0
timestamp           0
model_attr          0
category            0
brand          961834
year                0
user_attr     1118830
split               0
dtype: int64

Let us check the types of values in the missing values columns

"""for column brand"""
raw_electronics_db.brand.value_counts().head()

Bose        29222
Logitech    29203
Sony        20360
Mpow        17477
EldHus      16530
Name: brand, dtype: int64

This column pertains to brands.Some products can be of less recognised brands and hence their data might not be visible

So we replace missing values with "Other".

'''Replacing the NA values in (Gender) user_attr column with "NA"'''
raw_electronics_db.brand.fillna("Other", inplace=True)

"""for column user_attr"""
raw_electronics_db.user_attr.value_counts()

Male      87424
Female    86700
Name: user_attr, dtype: int64

Here we can replace the missing values with 'Not Applicable or NA' as this column pertains to gender.

Some products can be unisex or not specified as per gender.

'''Replacing the NA values in (Gender) user_attr column with "NA"'''
raw_electronics_db.user_attr.fillna("NA", inplace=True)

Rechecking for missing values

raw_electronics_db.isna().sum()

item_id       0
user_id       0
rating        0
timestamp     0
model_attr    0
category      0
brand         0
year          0
user_attr     0
split         0
dtype: int64

From looking above, we see that there are no missing values in Electronics database

---

Let us check missing values for Modcloth dataframe

raw_modcloth_db.isna().sum()

item_id           0
user_id           1
rating            0
timestamp         0
size          21760
fit           18506
user_attr      8367
model_attr        0
category          0
brand         73980
year              0
split             0
dtype: int64

"brand" column pertains to brands.Some products can be of less recognised brands and hence their data might not be visible.

So we replace missing values with "Other".

raw_modcloth_db.brand.fillna("Other", inplace=True)

"""We will not be making use of columns size, fit user_attr, so we will ignore those columns"""

'We will not be making use of columns size, fit user_attr, so we will ignore those columns'

3. Feature Engineering and Exploratory Data Analysis

For building a recommendation system, we will be focusing on columns:

• item(item_id)
• user(user_id)
• rating(rating)
• time when the item was bought(timestamp)
• category(category)
• brand(brand)

select_electronics_db = raw_electronics_db[
    ["item_id", "user_id", "rating", "timestamp", "category", "brand"]
]

select_modcloth_db = raw_modcloth_db[
    ["item_id", "user_id", "rating", "timestamp", "category", "brand"]
]

Let us merge both these dataframes to form a big consolidated dataframe for electronics as well as modcloth

"""Consolidated DataFrame"""
consolidated_db = select_electronics_db.append(select_modcloth_db)

"""Let us check the data types of the columns"""
consolidated_db.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 1392847 entries, 0 to 99892
Data columns (total 6 columns):
 #   Column     Non-Null Count    Dtype  
---  ------     --------------    -----  
 0   item_id    1392847 non-null  int64  
 1   user_id    1392846 non-null  object 
 2   rating     1392847 non-null  float64
 3   timestamp  1392847 non-null  object 
 4   category   1392847 non-null  object 
 5   brand      1392847 non-null  object 
dtypes: float64(1), int64(1), object(4)
memory usage: 74.4+ MB

"""from the above we see that timestamp is not in datetime datatype, hence we convert it to datetime"""
consolidated_db.timestamp = [str(i)[:10] for i in consolidated_db.timestamp.tolist()]

consolidated_db.timestamp = pd.to_datetime(consolidated_db.timestamp)

"""Let us form columns of Year and Month for our further analysis"""
consolidated_db.insert(4, "Month", pd.DatetimeIndex(consolidated_db.timestamp).month)
consolidated_db.insert(5, "Year", pd.DatetimeIndex(consolidated_db.timestamp).year)

consolidated_db.head()

	item_id	user_id	rating	timestamp	Month	Year	category	brand
0	0	0	5.0	1999-06-13	6	1999	Portable Audio & Video	Other
1	0	1	5.0	1999-06-14	6	1999	Portable Audio & Video	Other
2	0	2	3.0	1999-06-17	6	1999	Portable Audio & Video	Other
3	0	3	1.0	1999-07-01	7	1999	Portable Audio & Video	Other
4	0	4	2.0	1999-07-06	7	1999	Portable Audio & Video	Other

"""Let us check the period for which we have the data"""

consolidated_db.sort_values(by="timestamp")

	item_id	user_id	rating	timestamp	Month	Year	category	brand
0	0	0	5.0	1999-06-13	6	1999	Portable Audio & Video	Other
1	0	1	5.0	1999-06-14	6	1999	Portable Audio & Video	Other
2	0	2	3.0	1999-06-17	6	1999	Portable Audio & Video	Other
3	0	3	1.0	1999-07-01	7	1999	Portable Audio & Video	Other
4	0	4	2.0	1999-07-06	7	1999	Portable Audio & Video	Other
...	...	...	...	...	...	...	...	...
99881	47325	rachel	1.0	2019-06-26	6	2019	Outerwear	Other
99885	77949	rachel	5.0	2019-06-26	6	2019	Bottoms	Other
99891	71607	amy	3.0	2019-06-27	6	2019	Outerwear	Jack by BB Dakota
99890	67194	Janice	5.0	2019-06-27	6	2019	Dresses	Other
99892	119732	sarah	3.0	2019-06-29	6	2019	Dresses	Other

1392847 rows × 8 columns

From the above, we see that we have data from June 1999 till June 2019 i.e. roughly 20 years

"""For counting purpose, we will create a column called as count with element as 1"""
consolidated_db["count"] = 1

consolidated_db.head()

	item_id	user_id	rating	timestamp	Month	Year	category	brand	count
0	0	0	5.0	1999-06-13	6	1999	Portable Audio & Video	Other	1
1	0	1	5.0	1999-06-14	6	1999	Portable Audio & Video	Other	1
2	0	2	3.0	1999-06-17	6	1999	Portable Audio & Video	Other	1
3	0	3	1.0	1999-07-01	7	1999	Portable Audio & Video	Other	1
4	0	4	2.0	1999-07-06	7	1999	Portable Audio & Video	Other	1

"""Let us check how many times each item has been bought & the average rating per item"""

buy_count = consolidated_db.groupby(by="item_id", as_index=False).agg(
    {"count": pd.Series.sum}
)
buy_count.rename(columns={"count": "total_sold_qty"}, inplace=True)

buy_count.head()

	item_id	total_sold_qty
0	0	118
1	1	58
2	2	3404
3	3	155
4	4	11

"""Let us create a function to get category of an item"""

item_category = consolidated_db[["item_id","category"]].drop_duplicates()

def get_item_category(item_id):
    return item_category[item_category.item_id == f"{item_id}"].values[0][1]

"""Let us check the buy count trend over the years"""
time_df = consolidated_db.groupby(by="timestamp", as_index=False).agg(
    {"count": pd.Series.sum}
)

Data Visualisation:

"""We will check the number of sales over time"""

fig = px.line(data_frame=time_df, x="timestamp", y="count")
fig.update_layout(
    title_text="Number of purchases over time", xaxis_title="Time", yaxis_title="Count"
)
fig.show()

It appears the purchases spiked in the year 2016-17.

This maybe due to boom in the E-retail sector with competing E-retail companies giving heavy discounts to boost their sales compared to their competitors

"""Let us check the sales performance of the categories every month"""
monthly_category_sale = consolidated_db.groupby(
    by=["category", "Month"], as_index=False
).agg({"count": pd.Series.sum})

fig = px.line(data_frame=monthly_category_sale, x="Month", y="count", color="category")
fig.update_layout(title_text="Category Sales Performance over Months")
fig.show()

From the above graph, it appears that for most of the categories, sales are maximum in the month of January & December i.e. at the year end

Dealing with Bias

Now, when it comes to rating, there will be cases where bias comes in play, such as:

1. The producer himself/herself must have bought the product & given it 5 rating or

2. there might be some loyal customers who might have given the product a 5 rating out of feeling of loyalty

To counter that, we will only consider reviews of those items where the purchase of items has been above 50

We have already prepared dataframe which gives the total quantity sold by every item.

We will merge that dataframe with the consolidate_db dataframe

"""Merging the dataframes"""
consolidated_db2 = pd.merge(consolidated_db, buy_count, on="item_id")

"""Considering cases where the sale of item is above 50 units"""
consolidated_db2 = consolidated_db2[consolidated_db2.total_sold_qty > 50]

Let us check the categories for which we need to build the recommendation system for

list_category = list(consolidated_db2.category.unique())
consolidated_db2.category.value_counts()

Headphones                 341966
Computers & Accessories    294515
Camera & Photo             161712
Accessories & Supplies     152388
Portable Audio & Video     132821
Tops                        30522
Dresses                     30033
Car Electronics & GPS       29147
Television & Video          27707
Home Audio                  23034
Bottoms                     21960
Wearable Technology         13647
Security & Surveillance      9315
Outerwear                    6380
Name: category, dtype: int64

We will check how many major brands are there per category

brand_df = consolidated_db2.groupby(by=["category", "brand"], as_index=False).agg(
    {"count": pd.Series.sum}
)
brand_df["count"] = 1
brand_df = brand_df.groupby(by="category", as_index=False).agg({"count": pd.Series.sum})

category_check = brand_df.sort_values(by="count", ascending=False)

"""Since one of the brand names in Others, we can substract 1 from each category to get the accurate count of brand names"""
category_check["count"] = category_check["count"] - 1

category_check

	category	count
4	Computers & Accessories	28
2	Camera & Photo	24
6	Headphones	20
9	Portable Audio & Video	15
0	Accessories & Supplies	11
11	Television & Video	8
3	Car Electronics & GPS	7
5	Dresses	7
7	Home Audio	6
12	Tops	5
1	Bottoms	3
8	Outerwear	3
10	Security & Surveillance	3
13	Wearable Technology	1

From the above, we see that:

1. Computers & Accessories has maximum brands (28 major and other brands)
2. Wearable Technology has minimum brands (1 major and other brands)
3. Except for Security & Surveillance and Wearable Technology, all other brands have 5 or more types of brands

4. Recommending Items based on Ratings

"""Let us form database of products sold in the past 2 years"""
past_2_years = consolidated_db2[consolidated_db2.timestamp > "2016-12-31"]

past_2_years.head()

	item_id	user_id	rating	timestamp	Month	Year	category	brand	count	total_sold_qty
3339	2	919978	5.0	2017-01-04	1	2017	Headphones	Other	1	3404
3340	2	919979	2.0	2017-01-04	1	2017	Headphones	Other	1	3404
3341	2	922707	5.0	2017-01-07	1	2017	Headphones	Other	1	3404
3342	2	922979	5.0	2017-01-08	1	2017	Headphones	Other	1	3404
3343	2	922980	1.0	2017-01-08	1	2017	Headphones	Other	1	3404

"""over-all item dataframe (1999 to 2019)"""
ovr_item_df = consolidated_db2.groupby(
    by=["category", "item_id", "brand", "total_sold_qty"], as_index=False
).agg({"rating": pd.Series.mean})
ovr_item_df.rating = ovr_item_df.rating.round(2)
ovr_item_df = ovr_item_df.sort_values(by="rating", ascending=False)


"""last 2 years item dataframe (2017 to 2019)"""
ovr_2_item_df = past_2_years.groupby(
    by=["category", "item_id", "brand", "total_sold_qty"], as_index=False
).agg({"rating": pd.Series.mean})
ovr_2_item_df.rating = ovr_2_item_df.rating.round(2)
ovr_2_item_df = ovr_2_item_df.sort_values(by="rating", ascending=False)

Let us Focus on Overall Data Frame

Let us recommend products per category

def top5_item_rating(category_input):
    """this function gives upto top 5 (or less) items in a given category based on rating in descending order"""
    temp_top5_rating = ovr_item_df[ovr_item_df.category == category_input].head(5)
    return [
        category_input,
        temp_top5_rating.item_id.tolist(),
        temp_top5_rating.brand.tolist(),
        temp_top5_rating.rating.tolist(),
    ]

4.a. Item Recommendation as per ratings received from 1999 to 2019

print("Recommending Items based on their ratings received from 1999 to 2019:")
print("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n\n")
for i in range(14):
    item_list = top5_item_rating(list_category[i])
    df = pd.DataFrame(
        {"Product": item_list[1], "Brand": item_list[2], "Rating": item_list[3]}
    )
    print(f"For Category:{item_list[0]}")
    print()
    print(f"{df}")
    print("_______________________________")
    print()
    print()

Recommending Items based on their ratings received from 1999 to 2019:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


For Category:Portable Audio & Video

   Product  Brand  Rating
0     1647  Other    4.99
1     3214  Other    4.84
2     2667  Other    4.78
3     5802  Other    4.71
4     7715  Other    4.69
_______________________________


For Category:Computers & Accessories

   Product  Brand  Rating
0     4050  Other    4.90
1     4966  Plemo    4.90
2     5607  Other    4.84
3     7382  Other    4.83
4     9392  Other    4.82
_______________________________


For Category:Headphones

   Product   Brand  Rating
0     4576  EldHus    4.87
1     4413   Other    4.79
2     4320   Other    4.79
3     4013   Other    4.75
4     4584   Other    4.74
_______________________________


For Category:Camera & Photo

   Product     Brand  Rating
0     2513     Other    5.00
1     8484     Canon    4.91
2     7460     Other    4.88
3     4283  Fujifilm    4.88
4     4367     Other    4.88
_______________________________


For Category:Accessories & Supplies

   Product  Brand  Rating
0     3103  Other    4.89
1     7570  Other    4.88
2     3672  Other    4.88
3     7163  ebasy    4.87
4     5769  Other    4.85
_______________________________


For Category:Television & Video

   Product      Brand  Rating
0     5030      Other    4.72
1      476  Panasonic    4.69
2     9366      Other    4.66
3     8177      Other    4.66
4     2462      Other    4.56
_______________________________


For Category:Car Electronics & GPS

   Product  Brand  Rating
0     4659  Other    4.96
1     1991  Other    4.87
2     3498  Other    4.86
3     3300  Other    4.81
4     5559  Other    4.74
_______________________________


For Category:Home Audio

   Product  Brand  Rating
0     1546  Other    4.80
1     3443  Other    4.76
2     7384  Other    4.73
3      811  Other    4.73
4     1006  Other    4.70
_______________________________


For Category:Security & Surveillance

   Product  Brand  Rating
0     4027  Other    4.67
1     5031  Other    4.65
2     1840  Other    4.50
3     3754  Other    4.49
4     2687  Other    4.47
_______________________________


For Category:Wearable Technology

   Product  Brand  Rating
0     4968  Other    4.80
1     7759  Other    4.37
2     8752  Other    4.31
3     5033  Other    4.27
4     6438  Other    4.17
_______________________________


For Category:Dresses

   Product     Brand  Rating
0   154064  ModCloth    4.78
1   142983     Other    4.67
2   143686     Other    4.65
3   110169       Ryu    4.63
4   112559     Other    4.62
_______________________________


For Category:Outerwear

   Product         Brand  Rating
0   135555      ModCloth    4.61
1   123500      ModCloth    4.54
2    35272         Other    4.52
3   129113  Steve Madden    4.51
4    27439  Steve Madden    4.48
_______________________________


For Category:Bottoms

   Product            Brand  Rating
0   133539  Miss Candyfloss    4.76
1   144572            Other    4.74
2   153454         ModCloth    4.62
3   141880         ModCloth    4.61
4   148257         ModCloth    4.61
_______________________________


For Category:Tops

   Product           Brand  Rating
0    67022    Out of Print    4.79
1   143764  Daisey Natives    4.74
2    70230        Kin Ship    4.73
3   122266        Kin Ship    4.72
4   116736           Other    4.64
_______________________________

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

items_excel_df = pd.DataFrame(columns=["Recommendation_System_Type", "Category" , "Product", "Brand","Rating", "Sales"])

for i in range(14):
    item_list = top5_item_rating(list_category[i])
    df = pd.DataFrame(
        {"Recommendation_System_Type":"Overall Rating", "Category": item_list[0], "Product": item_list[1], "Brand": item_list[2], "Rating": item_list[3]}
    )

    items_excel_df = items_excel_df.append(df)

4.b. Brandwise Item Recommendation as per ratings received from 1999 to 2019

Let us recommend products as per brand

brand_df = (
    consolidated_db2.groupby(by=["category", "brand"], as_index=False)
    .agg({"rating": pd.Series.mean})
    .sort_values(by="rating", ascending=False)
)


"""We will not be considering the Other label in the brands column"""
brand_df = brand_df[brand_df.brand != "Other"]
brand_df.rating = brand_df.rating.round(2)

We will build a function which returns top 5 brands in every category and the top products for a particular category for a particular brand in descending order of their rating

def top5_brands_rating(category_name):
    """This function returns a list of top brands and their ratings upon entering the category"""
    temp_df_1 = brand_df[brand_df.category == category_name].head()
    return [category_name, temp_df_1.brand.tolist(), temp_df_1.rating.tolist()]

def top_products_based_on_brand(category_name, brand_name):
    """This function prints dataframe of items and their ratings upon entering the category and brand"""
    temp_df_2 = ovr_item_df[
        (ovr_item_df.category == category_name) & (ovr_item_df.brand == brand_name)
    ].head()
    item_list = temp_df_2.item_id.tolist()
    rating_list = temp_df_2.rating.tolist()
    df = pd.DataFrame({"Item_ID": item_list, "Rating": rating_list})
    print(df)

print("Recommending Items as per Top Brands from 1999 to 2019:")
print("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n\n")

for k in range(14):
    print(f"For Category: {list_category[k]}, top brands recommended are:\n")
    sol = top5_brands_rating(list_category[k])
    for j in range(len(sol[1])):
        brand_id = sol[1][j]
        brand_rating = sol[2][j]
        print(f"Brand Rank {j+1}: {brand_id}")
        print(f"Brand Rating: {brand_rating}")
        print()
        top_products_based_on_brand(list_category[k], brand_id)
        print("_______________________________")
        print("-------------------------------")
    print()
    print(
        "________________________________________________________________________________________________"
    )
    print(
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
    )
    print()

Recommending Items as per Top Brands from 1999 to 2019:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


For Category: Portable Audio & Video, top brands recommended are:

Brand Rank 1: Mpow
Brand Rating: 4.48

   Item_ID  Rating
0     5974    4.49
1     5748    4.47
_______________________________
-------------------------------
Brand Rank 2: Bose
Brand Rating: 4.43

   Item_ID  Rating
0     5171    4.51
1     5273    4.48
2     5156    4.47
3     7022    4.41
4     3195    4.37
_______________________________
-------------------------------
Brand Rank 3: Apple
Brand Rating: 4.35

   Item_ID  Rating
0     1139    4.35
_______________________________
-------------------------------
Brand Rank 4: DBPOWER
Brand Rating: 4.16

   Item_ID  Rating
0     4009    4.20
1     2864    4.00
2     4079    3.46
_______________________________
-------------------------------
Brand Rank 5: Logitech
Brand Rating: 4.07

   Item_ID  Rating
0     1948    4.07
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Computers & Accessories, top brands recommended are:

Brand Rank 1: Plemo
Brand Rating: 4.61

   Item_ID  Rating
0     4966    4.90
1     6839    4.76
2     4869    4.75
3     3258    4.75
4     3916    4.66
_______________________________
-------------------------------
Brand Rank 2: Cooper Cases
Brand Rating: 4.55

   Item_ID  Rating
0     8655    4.82
1     9089    4.54
2     9203    4.53
3     3818    4.18
_______________________________
-------------------------------
Brand Rank 3: Fintie
Brand Rating: 4.52

   Item_ID  Rating
0     5226    4.82
1     7812    4.78
2     4898    4.72
3     5245    4.71
4     5123    4.69
_______________________________
-------------------------------
Brand Rank 4: Apple
Brand Rating: 4.43

   Item_ID  Rating
0      385    4.70
1     5405    4.65
2     4302    4.52
3     6562    4.51
4     5557    4.46
_______________________________
-------------------------------
Brand Rank 5: EldHus
Brand Rating: 4.43

   Item_ID  Rating
0     1549    4.60
1     4369    4.50
2     2863    4.43
3     4144    4.07
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Headphones, top brands recommended are:

Brand Rank 1: Tiamat
Brand Rating: 4.43

   Item_ID  Rating
0     6853    4.66
1     3230    4.60
2     4708    4.51
3     7472    4.50
4     3557    4.43
_______________________________
-------------------------------
Brand Rank 2: ebasy
Brand Rating: 4.4

   Item_ID  Rating
0     5521    4.54
1     2072    4.28
_______________________________
-------------------------------
Brand Rank 3: Bose
Brand Rating: 4.21

   Item_ID  Rating
0     3432    4.62
1     2595    4.60
2     2469    4.57
3     1825    4.47
4     2480    4.46
_______________________________
-------------------------------
Brand Rank 4: TaoTronics
Brand Rating: 4.2

   Item_ID  Rating
0     7721    4.71
1     6437    4.43
2     5812    4.42
3     4446    4.32
4     4922    4.28
_______________________________
-------------------------------
Brand Rank 5: Sony
Brand Rating: 4.17

   Item_ID  Rating
0     6493    4.33
1     4959    4.24
2     3544    4.23
3     8196    4.23
4     3660    4.22
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Camera & Photo, top brands recommended are:

Brand Rank 1: Fintie
Brand Rating: 4.57

   Item_ID  Rating
0     4581    4.57
_______________________________
-------------------------------
Brand Rank 2: Savage
Brand Rating: 4.56

   Item_ID  Rating
0      672    4.81
1     1564    4.71
2     1416    4.69
3      466    4.58
4     1375    4.57
_______________________________
-------------------------------
Brand Rank 3: Sony
Brand Rating: 4.37

   Item_ID  Rating
0     2083    4.79
1       49    4.77
2     3738    4.75
3     1087    4.70
4     2657    4.65
_______________________________
-------------------------------
Brand Rank 4: Mpow
Brand Rating: 4.36

   Item_ID  Rating
0     8004    4.36
_______________________________
-------------------------------
Brand Rank 5: Neewer
Brand Rating: 4.33

   Item_ID  Rating
0     7443    4.57
1     7101    4.47
2     1605    4.45
3     3783    4.31
4     8341    4.16
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Accessories & Supplies, top brands recommended are:

Brand Rank 1: Sennheiser
Brand Rating: 4.6

   Item_ID  Rating
0     1560    4.69
1      227    4.41
2     4590    4.29
_______________________________
-------------------------------
Brand Rank 2: Etre Jeune
Brand Rating: 4.49

   Item_ID  Rating
0     7827    4.67
1     3229    4.58
2       79    4.51
3     5081    4.49
4     1473    4.34
_______________________________
-------------------------------
Brand Rank 3: Fintie
Brand Rating: 4.48

   Item_ID  Rating
0     3777    4.49
1     5806    4.48
_______________________________
-------------------------------
Brand Rank 4: ebasy
Brand Rating: 4.4

   Item_ID  Rating
0     7163    4.87
1     1274    4.61
2     4036    4.58
3     4821    4.28
4     3576    4.26
_______________________________
-------------------------------
Brand Rank 5: EldHus
Brand Rating: 4.21

   Item_ID  Rating
0     2106    4.75
1     5289    4.50
2     9188    4.49
3     8954    4.48
4     9169    4.46
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Television & Video, top brands recommended are:

Brand Rank 1: Bose
Brand Rating: 4.36

   Item_ID  Rating
0     5202    4.36
_______________________________
-------------------------------
Brand Rank 2: Panasonic
Brand Rating: 4.33

   Item_ID  Rating
0      476    4.69
1      371    4.53
2     1618    4.43
3     1326    4.40
4     1563    4.21
_______________________________
-------------------------------
Brand Rank 3: Toshiba
Brand Rating: 4.19

   Item_ID  Rating
0     4177    4.31
1     2214    4.09
2     2173    3.97
3     2984    3.89
_______________________________
-------------------------------
Brand Rank 4: Pyle
Brand Rating: 4.12

   Item_ID  Rating
0     2454    4.25
1     1984    3.60
_______________________________
-------------------------------
Brand Rank 5: Samsung
Brand Rating: 4.02

   Item_ID  Rating
0     2129    4.02
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Car Electronics & GPS, top brands recommended are:

Brand Rank 1: JVC
Brand Rating: 4.2

   Item_ID  Rating
0     1037    4.48
1      980    4.12
2      441    3.97
_______________________________
-------------------------------
Brand Rank 2: Garmin
Brand Rating: 4.17

   Item_ID  Rating
0     1323    4.43
1      283    4.42
2     6625    4.16
3     1105    4.09
4     1668    4.08
_______________________________
-------------------------------
Brand Rank 3: Sony
Brand Rating: 4.16

   Item_ID  Rating
0     2690    4.16
_______________________________
-------------------------------
Brand Rank 4: Mpow
Brand Rating: 3.97

   Item_ID  Rating
0     5749    3.97
_______________________________
-------------------------------
Brand Rank 5: Uniden
Brand Rating: 3.94

   Item_ID  Rating
0     1056    4.04
1     6731    3.17
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Home Audio, top brands recommended are:

Brand Rank 1: EldHus
Brand Rating: 4.51

   Item_ID  Rating
0     6662    4.51
_______________________________
-------------------------------
Brand Rank 2: Sony
Brand Rating: 4.49

   Item_ID  Rating
0      511    4.49
_______________________________
-------------------------------
Brand Rank 3: TaoTronics
Brand Rating: 4.17

   Item_ID  Rating
0     5054    4.17
1     8375    4.17
2     6328    4.16
_______________________________
-------------------------------
Brand Rank 4: Bose
Brand Rating: 4.04

   Item_ID  Rating
0      211    4.47
1     3625    4.30
2      178    2.53
_______________________________
-------------------------------
Brand Rank 5: Pyle
Brand Rating: 3.82

   Item_ID  Rating
0     1236    4.43
1     5096    4.27
2     5459    3.98
3     1205    3.96
4      827    3.89
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Security & Surveillance, top brands recommended are:

Brand Rank 1: Uniden
Brand Rating: 3.79

   Item_ID  Rating
0     3989    3.94
1     6638    3.75
2     1250    3.74
3     3603    3.19
_______________________________
-------------------------------
Brand Rank 2: Samsung
Brand Rating: 3.75

   Item_ID  Rating
0     3293    3.75
_______________________________
-------------------------------
Brand Rank 3: Funlux
Brand Rating: 3.16

   Item_ID  Rating
0     4831    3.16
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Wearable Technology, top brands recommended are:

Brand Rank 1: Sony
Brand Rating: 3.68

   Item_ID  Rating
0     5402    3.68
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Dresses, top brands recommended are:

Brand Rank 1: Banned
Brand Rating: 4.61

   Item_ID  Rating
0   145278    4.61
_______________________________
-------------------------------
Brand Rank 2: Ryu
Brand Rating: 4.45

   Item_ID  Rating
0   110169    4.63
1   104356    4.56
2    48629    4.37
_______________________________
-------------------------------
Brand Rank 3: Pepaloves
Brand Rating: 4.37

   Item_ID  Rating
0   153481    4.37
_______________________________
-------------------------------
Brand Rank 4: ModCloth
Brand Rating: 4.3

   Item_ID  Rating
0   154064    4.78
1    70538    4.47
2   141105    4.34
3   137270    4.25
4   153445    4.25
_______________________________
-------------------------------
Brand Rank 5: Retrolicious
Brand Rating: 4.3

   Item_ID  Rating
0   140020    4.51
1    28252    4.47
2   119314    4.46
3    40141    4.38
4    56780    4.32
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Outerwear, top brands recommended are:

Brand Rank 1: Steve Madden
Brand Rating: 4.43

   Item_ID  Rating
0   129113    4.51
1    27439    4.48
2   110824    4.28
3   110820    4.25
_______________________________
-------------------------------
Brand Rank 2: Jack by BB Dakota
Brand Rating: 4.12

   Item_ID  Rating
0    71607    4.12
_______________________________
-------------------------------
Brand Rank 3: ModCloth
Brand Rating: 4.07

   Item_ID  Rating
0   135555    4.61
1   123500    4.54
2   143136    4.34
3    85507    4.06
4    58990    3.93
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Bottoms, top brands recommended are:

Brand Rank 1: Miss Candyfloss
Brand Rating: 4.47

   Item_ID  Rating
0   133539    4.76
1   133399    4.44
2   121272    4.38
_______________________________
-------------------------------
Brand Rank 2: Pink Martini
Brand Rating: 4.41

   Item_ID  Rating
0   111781    4.41
_______________________________
-------------------------------
Brand Rank 3: ModCloth
Brand Rating: 4.27

   Item_ID  Rating
0   153454    4.62
1   141880    4.61
2   148257    4.61
3   153372    4.59
4   140727    4.59
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Tops, top brands recommended are:

Brand Rank 1: Daisey Natives
Brand Rating: 4.74

   Item_ID  Rating
0   143764    4.74
_______________________________
-------------------------------
Brand Rank 2: Kin Ship
Brand Rating: 4.72

   Item_ID  Rating
0    70230    4.73
1   122266    4.72
_______________________________
-------------------------------
Brand Rank 3: Out of Print
Brand Rating: 4.6

   Item_ID  Rating
0    67022    4.79
1   113643    4.25
_______________________________
-------------------------------
Brand Rank 4: ModCloth
Brand Rating: 4.41

   Item_ID  Rating
0   153726    4.57
1   118020    4.53
2    40899    4.52
3    59176    4.51
4   126993    4.37
_______________________________
-------------------------------
Brand Rank 5: Banned
Brand Rating: 3.9

   Item_ID  Rating
0   146244     3.9
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

def top_products_based_on_brand_2(category_name, brand_name):
    """This function prints dataframe of items and their ratings upon entering the category and brand"""
    temp_df_2 = ovr_item_df[
        (ovr_item_df.category == category_name) & (ovr_item_df.brand == brand_name)
    ].head()
    item_list = temp_df_2.item_id.tolist()
    rating_list = temp_df_2.rating.tolist()
    return [f"{item_list[i]} - {rating_list[i]}" for i in range(len(item_list))]


brand_excel_df = pd.DataFrame(columns=["Recommendation_System_Type", "Category" ,"Brand","Brand_Rating", "Top_Products", "Sales"])

for k in range(14):
    sol = top5_brands_rating(list_category[k])
    for j in range(len(sol[1])):
        brand_id = sol[1][j]
        brand_rating = sol[2][j]
        brand_excel_df.loc[len(brand_excel_df.index)] = ["Overall_Brandwise_Recommendation",list_category[k],brand_id,brand_rating,top_products_based_on_brand_2(list_category[k], brand_id), None]

Let us focus on trend for last 2 Years

def last2_item_suggestion(category_name):
    """this function gives list of top items, their respective brands and their ratings"""
    temp_df3 = ovr_2_item_df[ovr_2_item_df.category == category_name].head()
    return [
        category_name,
        temp_df3.item_id.tolist(),
        temp_df3.brand.tolist(),
        temp_df3.rating.tolist(),
    ]

4.c. Latest Trend: Item Recommendation as per ratings received over the last 2 years

print("Recommending Items based on their ratings received over the past 2 years:")
print("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n\n")
for i in range(14):
    item_list = last2_item_suggestion(list_category[i])
    df = pd.DataFrame(
        {"Product": item_list[1], "Brand": item_list[2], "Rating": item_list[3]}
    )
    print(f"For Category:{item_list[0]}")
    print()
    print(f"{df}")
    print("_______________________________")
    print()
    print()

Recommending Items based on their ratings received over the past 2 years:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


For Category:Portable Audio & Video

   Product   Brand  Rating
0     1064   Other     5.0
1      816  Uniden     5.0
2      790   Other     5.0
3      570  Archos     5.0
4     1149    Sony     5.0
_______________________________


For Category:Computers & Accessories

   Product  Brand  Rating
0     2939  Other     5.0
1     2958  Other     5.0
2      977  Apple     5.0
3      896  Other     5.0
4     7097  Other     5.0
_______________________________


For Category:Headphones

   Product  Brand  Rating
0     7222  Other     5.0
1     4320  Other     5.0
2     4458  Other     5.0
3     4140  Other     5.0
4     4568  Other     5.0
_______________________________


For Category:Camera & Photo

   Product      Brand  Rating
0     1598  ViewSonic     5.0
1      506      Other     5.0
2     6499      Other     5.0
3      520       Sony     5.0
4      537      Kodak     5.0
_______________________________


For Category:Accessories & Supplies

   Product  Brand  Rating
0     3960  Other     5.0
1     3661  Other     5.0
2     3719  Other     5.0
3     4112  Other     5.0
4     4469  Other     5.0
_______________________________


For Category:Television & Video

   Product    Brand  Rating
0     8122    Other     5.0
1     5030    Other     5.0
2     3656    Other     5.0
3     2606    Other     5.0
4     2214  Toshiba     5.0
_______________________________


For Category:Car Electronics & GPS

   Product  Brand  Rating
0     4465  Other     5.0
1     3300  Other     5.0
2     1521  Other     5.0
3     4476  Other     5.0
4     1401  Other     5.0
_______________________________


For Category:Home Audio

   Product  Brand  Rating
0     4481  Other     5.0
1     5636  Other     5.0
2     6198  Other     5.0
3     2655  Other     5.0
4      211   Bose     5.0
_______________________________


For Category:Security & Surveillance

   Product  Brand  Rating
0     1840  Other    5.00
1     5792  Other    4.54
2     3754  Other    4.42
3     2687  Other    4.33
4     4347  Other    4.29
_______________________________


For Category:Wearable Technology

   Product  Brand  Rating
0     4968  Other    4.50
1     8752  Other    4.29
2     2830  Other    4.26
3     6438  Other    4.10
4     8851  Other    4.02
_______________________________


For Category:Dresses

   Product           Brand  Rating
0   154064        ModCloth    4.78
1   146840           Other    4.72
2    61110  Chi Chi London    4.71
3   110169             Ryu    4.70
4   135791           Other    4.69
_______________________________


For Category:Outerwear

   Product         Brand  Rating
0   135555      ModCloth    4.60
1   123500      ModCloth    4.58
2   129113  Steve Madden    4.57
3    35272         Other    4.51
4    47325         Other    4.46
_______________________________


For Category:Bottoms

   Product            Brand  Rating
0   133539  Miss Candyfloss    4.83
1   102736            Other    4.81
2    70478            Other    4.71
3   114371            Other    4.69
4   137232         ModCloth    4.64
_______________________________


For Category:Tops

   Product           Brand  Rating
0    67022    Out of Print    4.81
1   122266        Kin Ship    4.78
2    63583           Other    4.73
3   137370           Other    4.73
4   143764  Daisey Natives    4.71
_______________________________

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

for i in range(14):
    item_list = last2_item_suggestion(list_category[i])
    df = pd.DataFrame(
        {"Recommendation_System_Type":"Latest Trend Rating", "Category": item_list[0], "Product": item_list[1], "Brand": item_list[2], "Rating": item_list[3]}
    )
    items_excel_df = items_excel_df.append(df)

items_excel_df.to_csv("item_recommendation.csv")

4.d. Latest Trend: Brandwise Item Recommendation as per ratings over the last 2 years

Let us recommend products as per brand

"""Forming the required dataframe for past 2 years keeping category as the main parameter"""
last2_brand_df = (
    past_2_years.groupby(by=["category", "brand"], as_index=False)
    .agg({"rating": pd.Series.mean})
    .sort_values(by="rating", ascending=False)
)
last2_brand_df = last2_brand_df[last2_brand_df.brand != "Other"]
last2_brand_df.rating = last2_brand_df.rating.round(2)

def brand_top_items_2years(category_name, brand_name):
    """this function prints the dataframe of Items and their ratings based upon category and brand"""
    brand_items_2years = ovr_2_item_df[
        (ovr_2_item_df.brand == brand_name) & (ovr_2_item_df.category == category_name)
    ].head()
    brand_items = brand_items_2years.item_id.tolist()
    brand_items_rating = brand_items_2years.rating.tolist()
    df = pd.DataFrame({"Item_Code": brand_items, "Rating": brand_items_rating})
    print(df)

def last2_brands_suggestion(category_name):
    """this function gives top brand names and their ratings"""
    temp_df_1 = last2_brand_df[last2_brand_df.category == category_name].head()
    top_brand_names = temp_df_1.brand.tolist()
    top_brand_ratings = temp_df_1.rating.tolist()
    return [top_brand_names, top_brand_ratings]

print("Recommending Items as per Top Brands:")
print("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n\n")

for k in range(14):
    print(f"For Category: {list_category[k]}, top brands recommended are:\n")
    sol = last2_brands_suggestion(list_category[k])
    for j in range(len(sol[0])):
        brand_id = sol[0][j]
        brand_rating = sol[1][j]
        print(f"Brand Rank {j+1}: {brand_id}")
        print(f"Brand Rating: {brand_rating}")
        print()
        brand_top_items_2years(list_category[k], brand_id)
        print("_______________________________")
        print("-------------------------------")
    print()
    print(
        "________________________________________________________________________________________________"
    )
    print(
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
    )
    print()

Recommending Items as per Top Brands:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


For Category: Portable Audio & Video, top brands recommended are:

Brand Rank 1: Mpow
Brand Rating: 4.5

   Item_Code  Rating
0       5748    4.51
1       5974    3.33
_______________________________
-------------------------------
Brand Rank 2: Bose
Brand Rating: 4.27

   Item_Code  Rating
0       7022    4.36
1       5171    3.99
2       3195    3.95
3       5156    3.86
4       7049    3.84
_______________________________
-------------------------------
Brand Rank 3: DBPOWER
Brand Rating: 4.23

   Item_Code  Rating
0       4009    4.25
1       2864    4.00
2       4079    4.00
_______________________________
-------------------------------
Brand Rank 4: Sony
Brand Rating: 4.14

   Item_Code  Rating
0       1149    5.00
1       1154    5.00
2        332    5.00
3       1764    5.00
4       1131    4.33
_______________________________
-------------------------------
Brand Rank 5: JLAB
Brand Rating: 4.09

   Item_Code  Rating
0       2571    4.09
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Computers & Accessories, top brands recommended are:

Brand Rank 1: Fujifilm
Brand Rating: 5.0

   Item_Code  Rating
0       1432     5.0
_______________________________
-------------------------------
Brand Rank 2: Jabra
Brand Rating: 4.73

   Item_Code  Rating
0       2253    4.80
1       1722    4.62
2       2903    4.15
_______________________________
-------------------------------
Brand Rank 3: Fintie
Brand Rating: 4.61

   Item_Code  Rating
0       5245    5.00
1       5226    5.00
2       4696    5.00
3       4877    4.88
4       6123    4.80
_______________________________
-------------------------------
Brand Rank 4: Cooper Cases
Brand Rating: 4.6

   Item_Code  Rating
0       8655    4.84
1       9089    4.58
2       9203    4.57
3       3818    4.27
_______________________________
-------------------------------
Brand Rank 5: Plemo
Brand Rating: 4.43

   Item_Code  Rating
0       6839    4.51
1       3916    4.50
2       5989    4.45
3       3556    4.20
4       5168    3.00
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Headphones, top brands recommended are:

Brand Rank 1: ebasy
Brand Rating: 4.4

   Item_Code  Rating
0       5521    4.54
1       2072    4.30
_______________________________
-------------------------------
Brand Rank 2: Mpow
Brand Rating: 4.16

   Item_Code  Rating
0       4019    5.00
1       6819    4.30
2       7809    4.08
3       6689    3.95
4       6835    3.86
_______________________________
-------------------------------
Brand Rank 3: Philips
Brand Rating: 4.11

   Item_Code  Rating
0        547    5.00
1       1499    4.50
2       5257    4.48
3       1466    4.33
4       7526    4.26
_______________________________
-------------------------------
Brand Rank 4: Kensington
Brand Rating: 4.07

   Item_Code  Rating
0        923    4.07
_______________________________
-------------------------------
Brand Rank 5: Bose
Brand Rating: 4.01

   Item_Code  Rating
0       2469    5.00
1       1806    5.00
2        338    5.00
3       1447    5.00
4       1860    4.75
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Camera & Photo, top brands recommended are:

Brand Rank 1: JVC
Brand Rating: 4.8

   Item_Code  Rating
0        465     4.8
_______________________________
-------------------------------
Brand Rank 2: Fintie
Brand Rating: 4.68

   Item_Code  Rating
0       4581    4.68
_______________________________
-------------------------------
Brand Rank 3: Sony
Brand Rating: 4.63

   Item_Code  Rating
0        520    5.00
1        795    5.00
2         49    5.00
3       2491    5.00
4       2083    4.89
_______________________________
-------------------------------
Brand Rank 4: Canon
Brand Rating: 4.59

   Item_Code  Rating
0        725    5.00
1        698    5.00
2        143    5.00
3       2979    5.00
4       9189    4.89
_______________________________
-------------------------------
Brand Rank 5: Savage
Brand Rating: 4.43

   Item_Code  Rating
0        672    4.91
1       1564    4.68
2       2556    4.62
3       1375    4.62
4       1263    4.39
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Accessories & Supplies, top brands recommended are:

Brand Rank 1: Fintie
Brand Rating: 4.38

   Item_Code  Rating
0       3777    4.68
1       5806    4.25
_______________________________
-------------------------------
Brand Rank 2: Sennheiser
Brand Rating: 4.38

   Item_Code  Rating
0       1560    4.65
1       4590    3.82
_______________________________
-------------------------------
Brand Rank 3: Pyle
Brand Rating: 4.32

   Item_Code  Rating
0       4225    4.38
1       1744    4.11
2       5258    4.00
_______________________________
-------------------------------
Brand Rank 4: ebasy
Brand Rating: 4.31

   Item_Code  Rating
0       3576    4.67
1       1274    4.50
2       4821    4.29
_______________________________
-------------------------------
Brand Rank 5: Etre Jeune
Brand Rating: 4.26

   Item_Code  Rating
0         79    4.71
1       7827    4.55
2       8466    4.30
3       5081    4.24
4       1473    4.13
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Television & Video, top brands recommended are:

Brand Rank 1: Panasonic
Brand Rating: 4.42

   Item_Code  Rating
0       1618    5.00
1       1563    5.00
2       1326    5.00
3       1296    4.67
4       1604    2.50
_______________________________
-------------------------------
Brand Rank 2: Bose
Brand Rating: 4.34

   Item_Code  Rating
0       5202    4.34
_______________________________
-------------------------------
Brand Rank 3: Pyle
Brand Rating: 3.82

   Item_Code  Rating
0       1984    3.82
_______________________________
-------------------------------
Brand Rank 4: DBPOWER
Brand Rating: 3.06

   Item_Code  Rating
0       8074    3.24
1       4633    2.50
2       2654    2.50
_______________________________
-------------------------------
Brand Rank 5: Samsung
Brand Rating: 3.03

   Item_Code  Rating
0       2129    3.03
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Car Electronics & GPS, top brands recommended are:

Brand Rank 1: JVC
Brand Rating: 4.33

   Item_Code  Rating
0       1037    4.33
_______________________________
-------------------------------
Brand Rank 2: Uniden
Brand Rating: 3.85

   Item_Code  Rating
0       1056    4.00
1       6731    2.75
_______________________________
-------------------------------
Brand Rank 3: Garmin
Brand Rating: 3.78

   Item_Code  Rating
0       1323    4.60
1        283    4.35
2       6625    4.05
3       1105    3.99
4       7936    3.97
_______________________________
-------------------------------
Brand Rank 4: Pyle
Brand Rating: 3.56

   Item_Code  Rating
0        593    3.95
1       2431    3.83
2       8335    3.75
3        638    3.56
4       8041    3.44
_______________________________
-------------------------------
Brand Rank 5: Mpow
Brand Rating: 3.5

   Item_Code  Rating
0       5749     3.5
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Home Audio, top brands recommended are:

Brand Rank 1: Sony
Brand Rating: 4.54

   Item_Code  Rating
0        511    4.54
_______________________________
-------------------------------
Brand Rank 2: EldHus
Brand Rating: 4.43

   Item_Code  Rating
0       6662    4.43
_______________________________
-------------------------------
Brand Rank 3: TaoTronics
Brand Rating: 4.08

   Item_Code  Rating
0       8375    4.16
1       5054    4.02
2       6328    3.88
_______________________________
-------------------------------
Brand Rank 4: Pyle
Brand Rating: 3.73

   Item_Code  Rating
0       2495    4.40
1       5096    4.37
2        827    4.22
3       1236    4.16
4       1205    4.06
_______________________________
-------------------------------
Brand Rank 5: Bose
Brand Rating: 3.56

   Item_Code  Rating
0        211     5.0
1       3625     3.0
2        178     2.5
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Security & Surveillance, top brands recommended are:

Brand Rank 1: Uniden
Brand Rating: 3.66

   Item_Code  Rating
0       3989    3.71
1       1250    3.67
2       3603    3.50
3       6638    3.48
_______________________________
-------------------------------
Brand Rank 2: Samsung
Brand Rating: 3.5

   Item_Code  Rating
0       3293     3.5
_______________________________
-------------------------------
Brand Rank 3: Funlux
Brand Rating: 1.57

   Item_Code  Rating
0       4831    1.57
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Wearable Technology, top brands recommended are:

Brand Rank 1: Sony
Brand Rating: 3.66

   Item_Code  Rating
0       5402    3.66
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Dresses, top brands recommended are:

Brand Rank 1: Banned
Brand Rating: 4.61

   Item_Code  Rating
0     145278    4.61
_______________________________
-------------------------------
Brand Rank 2: Ryu
Brand Rating: 4.55

   Item_Code  Rating
0     110169    4.70
1     104356    4.54
2      48629    4.50
_______________________________
-------------------------------
Brand Rank 3: Retrolicious
Brand Rating: 4.38

   Item_Code  Rating
0     140020    4.65
1      28252    4.47
2     119314    4.40
3      40141    4.38
4      56780    4.29
_______________________________
-------------------------------
Brand Rank 4: Pepaloves
Brand Rating: 4.37

   Item_Code  Rating
0     153481    4.37
_______________________________
-------------------------------
Brand Rank 5: Chi Chi London
Brand Rating: 4.2

   Item_Code  Rating
0      61110    4.71
1     129160    4.41
2     107539    4.39
3     141853    4.29
4     135364    4.25
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Outerwear, top brands recommended are:

Brand Rank 1: Steve Madden
Brand Rating: 4.22

   Item_Code  Rating
0     129113    4.57
1      27439    4.25
2     110820    4.16
3     110824    3.54
_______________________________
-------------------------------
Brand Rank 2: ModCloth
Brand Rating: 4.22

   Item_Code  Rating
0     135555    4.60
1     123500    4.58
2     143136    4.34
3      78510    4.12
4      56289    4.12
_______________________________
-------------------------------
Brand Rank 3: Jack by BB Dakota
Brand Rating: 4.02

   Item_Code  Rating
0      71607    4.02
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Bottoms, top brands recommended are:

Brand Rank 1: Pink Martini
Brand Rating: 4.46

   Item_Code  Rating
0     111781    4.46
_______________________________
-------------------------------
Brand Rank 2: Miss Candyfloss
Brand Rating: 4.42

   Item_Code  Rating
0     133539    4.83
1     133399    4.40
2     121272    4.25
_______________________________
-------------------------------
Brand Rank 3: ModCloth
Brand Rating: 4.38

   Item_Code  Rating
0     137232    4.64
1     153454    4.62
2     141880    4.62
3     148257    4.61
4     153372    4.59
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Tops, top brands recommended are:

Brand Rank 1: Kin Ship
Brand Rating: 4.76

   Item_Code  Rating
0     122266    4.78
1      70230    4.67
_______________________________
-------------------------------
Brand Rank 2: Daisey Natives
Brand Rating: 4.71

   Item_Code  Rating
0     143764    4.71
_______________________________
-------------------------------
Brand Rank 3: Out of Print
Brand Rating: 4.58

   Item_Code  Rating
0      67022    4.81
1     113643    4.00
_______________________________
-------------------------------
Brand Rank 4: ModCloth
Brand Rating: 4.3

   Item_Code  Rating
0      40899    4.63
1      59176    4.57
2     153726    4.57
3     118020    4.51
4     126993    4.40
_______________________________
-------------------------------
Brand Rank 5: Banned
Brand Rating: 4.02

   Item_Code  Rating
0     146244    4.02
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

def brand_top_items_2years_2(category_name, brand_name):
    """this function prints the dataframe of Items and their ratings based upon category and brand"""
    brand_items_2years = ovr_2_item_df[
        (ovr_2_item_df.brand == brand_name) & (ovr_2_item_df.category == category_name)
    ].head()
    brand_items = brand_items_2years.item_id.tolist()
    brand_items_rating = brand_items_2years.rating.tolist()
    return [f"{brand_items[i]} - {brand_items_rating[i]}" for i in range(len(brand_items))]



# brand_excel_df = pd.DataFrame(columns=["Recommendation_System_Type", "Category" ,"Brand","Brand_Rating", "Top_Products", "Sales"])

for k in range(14):
    sol = last2_brands_suggestion(list_category[k])
    for j in range(len(sol[0])):
        brand_id = sol[0][j]
        brand_rating = sol[1][j]
        brand_excel_df.loc[len(brand_excel_df.index)] = ["Latest_Brandwise_Recommendation",list_category[k],brand_id,brand_rating,brand_top_items_2years_2(list_category[k], brand_id), None]
        
        
brand_excel_df.to_csv("brand_recommendation.csv")

4.e. Collaborative Filtering Recommendation System

"""For Collaborative Filtering Recommendation System, we also consider the Users.
Hence, we will consider only those items rated by user who has also reviewed at least one more item 
(i.e. we will remove all users who have bought only 1 item)"""


"""let us form DataFrame of Users and the items they have purchased"""
rough = consolidated_db2["user_id"].value_counts().sort_values()
user_purchases = pd.DataFrame({"user_id": rough.index, "items_bought": rough.values})

# For demonstration purposes
user_purchases.tail()

	user_id	items_bought
1112962	Jennifer	154
1112963	Amanda	158
1112964	Jessica	159
1112965	Laura	160
1112966	Sarah	187

"""Let us form DataFrame of users who have bought more than 1 item"""
collab_df = consolidated_db2.copy()
collab_df = pd.merge(collab_df, user_purchases, on="user_id")
collab_df = collab_df[collab_df.items_bought > 1]


"""We will consider only relevant columns required for Collaborative Filtering in our DataFrame"""
collab_df = collab_df[["item_id", "user_id", "rating", "category"]]
collab_df.head()

	item_id	user_id	rating	category
7	0	7	2.0	Portable Audio & Video
8	2205	7	5.0	Computers & Accessories
9	9046	7	5.0	Accessories & Supplies
28	0	28	2.0	Portable Audio & Video
29	651	28	4.0	Camera & Photo

"""create pivot table"""
collab_df_pivot = collab_df.pivot(
    index="item_id", columns="user_id", values="rating"
).fillna(0)

"""create sparse matrix to pass to KNN model for training"""
features_df_matrix = csr_matrix(collab_df_pivot.values)

"""creating list of modcloth item_ids"""
item_list = collab_df_pivot.index.tolist()

"""creating KNN model"""
model_knn = NearestNeighbors(metric="cosine", algorithm="brute")

"""fitting the model"""
model_knn.fit(features_df_matrix)


"""Let us build a function which gives 5 nearest neighbours of item"""


def recommender_whole(item_id):
    query_index = np.random.choice(collab_df_pivot.shape[0])
    distances, indices = model_knn.kneighbors(
        collab_df_pivot[collab_df_pivot.index == item_id].values.reshape(1, -1),
        n_neighbors=6,
    )
    nearest_items = [collab_df_pivot.index[i] for i in indices.flatten()[1:]]
    nearest_items_category = [
        item_category[item_category.item_id == item]["category"].values[0]
        for item in nearest_items
    ]
    recommended_items_df = pd.DataFrame(
        {"Recommended_Item": nearest_items, "Category": nearest_items_category}
    )
    return recommended_items_df

for item in [np.random.choice(collab_df_pivot.index) for i in range(5)]:
    print()
    cat = item_category[item_category.item_id == 0]["category"].values[0]
    print(
        f"For Item_ID: {item} belonging to Category: {cat}, the top 5 items recommended are:"
    )
    print()
    print(recommender_whole(item))
    print()
    print(
        "_____________________________________________________________________________________________"
    )
    print(
        "============================================================================================="
    )

For Item_ID: 2037 belonging to Category: Portable Audio & Video, the top 5 items recommended are:

   Recommended_Item                 Category
0              2157               Headphones
1              5759               Headphones
2              6276  Computers & Accessories
3              3744   Accessories & Supplies
4              8072           Camera & Photo

_____________________________________________________________________________________________
=============================================================================================

For Item_ID: 2221 belonging to Category: Portable Audio & Video, the top 5 items recommended are:

   Recommended_Item                 Category
0               874           Camera & Photo
1              6372  Computers & Accessories
2              4040           Camera & Photo
3              9142               Headphones
4              5288  Computers & Accessories

_____________________________________________________________________________________________
=============================================================================================

For Item_ID: 3772 belonging to Category: Portable Audio & Video, the top 5 items recommended are:

   Recommended_Item                Category
0              4319  Portable Audio & Video
1              4385          Camera & Photo
2              4401              Headphones
3               167          Camera & Photo
4              3570              Headphones

_____________________________________________________________________________________________
=============================================================================================

For Item_ID: 5009 belonging to Category: Portable Audio & Video, the top 5 items recommended are:

   Recommended_Item               Category
0              7590             Headphones
1              5749  Car Electronics & GPS
2              2857             Headphones
3              3230             Headphones
4              9415             Headphones

_____________________________________________________________________________________________
=============================================================================================

For Item_ID: 5746 belonging to Category: Portable Audio & Video, the top 5 items recommended are:

   Recommended_Item                 Category
0              2502  Computers & Accessories
1              6797               Headphones
2              6254           Camera & Photo
3              5691               Headphones
4              7494               Headphones

_____________________________________________________________________________________________
=============================================================================================

Now that we have completed recommending items based on the ratings, we can also recommend items based upon the sales.

We can call these items as Best Selling Items

5. Recommending Items based on Demand of Items/Number of Units Sold (i.e. Best Selling items)

Let us group the items based on the units sold

best_selling_db = consolidated_db2[
    ["item_id", "timestamp", "category", "brand", "count"]
]

"""itemwise dataframe"""
best_selling_item = (
    best_selling_db.groupby(by=["item_id", "category", "brand"], as_index=False)
    .agg({"count": pd.Series.sum})
    .sort_values(by="count", ascending=False)
)

5.a. Best Selling Items from 1999 to 2019

def top5_bestselling_items_over(category_name):
    """this function prints the dataframe which shows Items, their respective brand and units sold for category entered"""
    df = best_selling_item[best_selling_item.category == category_name].head()
    item_list = df.item_id.tolist()
    brand_list = df.brand.tolist()
    sold_list = df["count"].tolist()
    temp_df = pd.DataFrame(
        {"Item_ID": item_list, "Brand": brand_list, "Sold_Units": sold_list}
    )
    print(temp_df)

"""brandwise dataframe"""

"""For suggesting brand, we cannot suggest "Other" as brand as we have substitued "Other" for Missing values.
Hence, we will drop "Other" brand and consider the rest"""
best_selling_brand_raw = best_selling_db[best_selling_db.brand != "Other"]

"""Forming DataFrame"""
best_selling_brand = (
    best_selling_brand_raw.groupby(by=["category", "brand"], as_index=False)
    .agg({"count": pd.Series.sum})
    .sort_values(by="count", ascending=False)
)

def best_selling_brand_ovr(category):
    """this function returns list of top brands and the units sold for the respective brands"""
    temp_df = best_selling_brand[best_selling_brand.category == category].head()
    brand_list = temp_df.brand.tolist()
    sold_list = temp_df["count"].tolist()
    return [brand_list, sold_list]

def bestselling_brand_items_ovr(category_name, brand_name):
    """this function prints the dataframe showing items and the respective units sold for those items for a given category"""
    temp_df = best_selling_item[
        (best_selling_item.brand == brand_name)
        & (best_selling_item.category == category_name)
    ].head()
    item_list = temp_df.item_id.tolist()
    sold_list = temp_df["count"].tolist()
    df = pd.DataFrame({"Item_ID": item_list, "Units Sold": sold_list})
    print(df)

print("Best Selling Items from 1999 to 2019:")
print("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n\n")
for i in range(len(list_category)):
    print(f"For Category:{list_category[i]}")
    print()
    top5_bestselling_items_over(list_category[i])
    print("_______________________________")
    print()
    print()

Best Selling Items from 1999 to 2019:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


For Category:Portable Audio & Video

   Item_ID    Brand  Sold_Units
0     7780    Other       28530
1     5063    Other        3875
2     5940    Other        3437
3     4009  DBPOWER        3340
4     7022     Bose        3183
_______________________________


For Category:Computers & Accessories

   Item_ID     Brand  Sold_Units
0     2031  Logitech        9393
1     2279     Other        6278
2     2486     Other        5810
3     1124  Logitech        5353
4      647  Logitech        4520
_______________________________


For Category:Headphones

   Item_ID  Brand  Sold_Units
0     1575   Sony        8622
1     2340  Other        5790
2     5795  Other        5554
3     1886  Other        5323
4     6210  Other        4916
_______________________________


For Category:Camera & Photo

   Item_ID     Brand  Sold_Units
0     5897     Other        3824
1     2237     Other        3777
2     1605    Neewer        3677
3     1875     Other        2636
4     6726  Polaroid        2086
_______________________________


For Category:Accessories & Supplies

   Item_ID  Brand  Sold_Units
0     1715  Other        5482
1     2483  Other        4325
2     6818  Other        4105
3     3995  Other        3716
4      791  Other        3316
_______________________________


For Category:Television & Video

   Item_ID    Brand  Sold_Units
0     2129  Samsung        4022
1     3617    Other        2420
2     4495    Other        1915
3     1735    Other        1252
4     6358    Other         991
_______________________________


For Category:Car Electronics & GPS

   Item_ID   Brand  Sold_Units
0     3728   Other        4345
1     6005   Other        2864
2      283  Garmin        2863
3     1105  Garmin        1457
4     5838   Other         938
_______________________________


For Category:Home Audio

   Item_ID       Brand  Sold_Units
0     5054  TaoTronics        4033
1     8375  TaoTronics        1252
2     3303       Other        1025
3     6328  TaoTronics         931
4     8802       Other         918
_______________________________


For Category:Security & Surveillance

   Item_ID   Brand  Sold_Units
0     9127   Other         994
1     6427   Other         572
2     7697   Other         536
3     3989  Uniden         529
4     3013   Other         523
_______________________________


For Category:Wearable Technology

   Item_ID  Brand  Sold_Units
0     2957  Other        3935
1     4968  Other        1300
2     7251  Other        1081
3     8851  Other         690
4     5402   Sony         637
_______________________________


For Category:Dresses

   Item_ID  Brand  Sold_Units
0    32405  Other        1599
1    32406  Other        1494
2    32403  Other        1378
3    16411  Other        1260
4    55967  Other        1164
_______________________________


For Category:Outerwear

   Item_ID         Brand  Sold_Units
0    11960         Other         854
1    27439  Steve Madden         772
2    32236         Other         751
3    35272         Other         561
4    58990      ModCloth         418
_______________________________


For Category:Bottoms

   Item_ID     Brand  Sold_Units
0    21296  ModCloth        1636
1    50818     Other         998
2    62447     Other         998
3    80238     Other         925
4    54222  ModCloth         749
_______________________________


For Category:Tops

   Item_ID     Brand  Sold_Units
0    34935     Other        1887
1    67507     Other        1357
2    57369     Other        1133
3    40899  ModCloth         957
4    59176  ModCloth         933
_______________________________

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

# def top5_bestselling_items_over_1(category_name):
#     """this function prints the dataframe which shows Items, their respective brand and units sold for category entered"""
#     df = best_selling_item[best_selling_item.category == category_name].head()
#     item_list = df.item_id.tolist()
#     brand_list = df.brand.tolist()
#     sold_list = df["count"].tolist()
#     temp_df = pd.DataFrame(
#         {"Item_ID": item_list, "Brand": brand_list, "Sold_Units": sold_list}
#     )
#     return temp_df

# temp_df2 = pd.DataFrame(columns=["Category", "Item_ID", "Brand", "Sold_Units"])

# for i in range(len(list_category)):
# #     print(f"For Category:{list_category[i]}")
#     temp_df = top5_bestselling_items_over_1(list_category[i])
#     temp_df.insert(0,"Category",list_category[i])
#     temp_df2 = temp_df2.append(temp_df)
    
# temp_df2.insert(0,"Recommendation_Type","Overall_Bestselling_Items")

# bestselling_item_recommendation = temp_df2.copy()

'Below code is to create the website for demonstration of Recommendation System, Hence it is commented out'

5.b. Brandwise Best Selling Items from 1999 to 2019

print("Recommending Items as per Top Selling Brand from 1999 to 2019:")
print("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n\n")

for k in range(len(list_category)):
    print(f"For Category: {list_category[k]}, best selling brands recommended are:\n")
    sol = best_selling_brand_ovr(list_category[k])
    for j in range(len(sol[0])):
        brand_id = sol[0][j]
        brand_rating = sol[1][j]
        print(f"Brand Number {j+1}: {brand_id}")
        print(f"Brand Units Sold: {brand_rating}")
        print()
        bestselling_brand_items_ovr(list_category[k], brand_id)
        print("_______________________________")
        print("-------------------------------")
    print()
    print(
        "________________________________________________________________________________________________"
    )
    print(
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
    )
    print()

Recommending Items as per Top Selling Brand from 1999 to 2019:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


For Category: Portable Audio & Video, best selling brands recommended are:

Brand Number 1: Bose
Brand Units Sold: 6826

   Item_ID  Units Sold
0     7022        3183
1     5171        1151
2     5273        1147
3     5156         630
4     3195         550
_______________________________
-------------------------------
Brand Number 2: DBPOWER
Brand Units Sold: 3812

   Item_ID  Units Sold
0     4009        3340
1     2864         385
2     4079          87
_______________________________
-------------------------------
Brand Number 3: Pyle
Brand Units Sold: 3083

   Item_ID  Units Sold
0       33         567
1     3468         293
2     4753         257
3     5445         244
4     3502         235
_______________________________
-------------------------------
Brand Number 4: Uniden
Brand Units Sold: 2802

   Item_ID  Units Sold
0     2223         806
1     2159         753
2     2434         275
3     4419         224
4     1848         199
_______________________________
-------------------------------
Brand Number 5: Mpow
Brand Units Sold: 1807

   Item_ID  Units Sold
0     5748        1754
1     5974          53
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Computers & Accessories, best selling brands recommended are:

Brand Number 1: Logitech
Brand Units Sold: 28699

   Item_ID  Units Sold
0     2031        9393
1     1124        5353
2      647        4520
3     1780        2711
4     1661        1387
_______________________________
-------------------------------
Brand Number 2: Linksys
Brand Units Sold: 7885

   Item_ID  Units Sold
0      710        1191
1     1055         866
2      679         847
3      142         731
4     1367         594
_______________________________
-------------------------------
Brand Number 3: Fintie
Brand Units Sold: 4348

   Item_ID  Units Sold
0     8736        1389
1     5354         413
2     3643         371
3     4696         183
4     2983         177
_______________________________
-------------------------------
Brand Number 4: Apple
Brand Units Sold: 4281

   Item_ID  Units Sold
0     5405        1020
1     4302         834
2     5364         737
3     6562         457
4     5329         404
_______________________________
-------------------------------
Brand Number 5: Bose
Brand Units Sold: 2999

   Item_ID  Units Sold
0     3106        2999
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Headphones, best selling brands recommended are:

Brand Number 1: Bose
Brand Units Sold: 18227

   Item_ID  Units Sold
0     5041        3452
1     1670        2969
2     4969        1675
3     7828        1532
4     7869        1336
_______________________________
-------------------------------
Brand Number 2: Mpow
Brand Units Sold: 14819

   Item_ID  Units Sold
0     6819        4058
1     6689        3967
2     5009        1523
3     7356        1026
4     6687         961
_______________________________
-------------------------------
Brand Number 3: Sennheiser
Brand Units Sold: 11406

   Item_ID  Units Sold
0      162        4002
1      259        3299
2     1242         964
3      554         555
4     1911         414
_______________________________
-------------------------------
Brand Number 4: Sony
Brand Units Sold: 11213

   Item_ID  Units Sold
0     1575        8622
1     3092         499
2     2450         296
3     7202         285
4     3480         258
_______________________________
-------------------------------
Brand Number 5: Skullcandy
Brand Units Sold: 9804

   Item_ID  Units Sold
0     2355        3347
1     5178        2366
2     8195        1222
3     3524         849
4     8350         471
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Camera & Photo, best selling brands recommended are:

Brand Number 1: Fujifilm
Brand Units Sold: 10882

   Item_ID  Units Sold
0     1258        1773
1       17        1073
2     3384         935
3     4407         744
4      751         660
_______________________________
-------------------------------
Brand Number 2: Kodak
Brand Units Sold: 7620

   Item_ID  Units Sold
0     1423        1352
1     1119         842
2     1289         589
3     1690         411
4      217         208
_______________________________
-------------------------------
Brand Number 3: Neewer
Brand Units Sold: 6261

   Item_ID  Units Sold
0     1605        3677
1     8341        1202
2     8077         481
3     5209         237
4     3783         185
_______________________________
-------------------------------
Brand Number 4: Nikon
Brand Units Sold: 6036

   Item_ID  Units Sold
0     1307         665
1     1156         484
2      296         452
3     1304         425
4     1282         381
_______________________________
-------------------------------
Brand Number 5: Canon
Brand Units Sold: 5849

   Item_ID  Units Sold
0     1593        1025
1      220         724
2     2881         604
3     3364         340
4     6434         279
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Accessories & Supplies, best selling brands recommended are:

Brand Number 1: EldHus
Brand Units Sold: 8113

   Item_ID  Units Sold
0     1861        1448
1     3539        1187
2      927        1185
3     1210         723
4     8282         661
_______________________________
-------------------------------
Brand Number 2: Etre Jeune
Brand Units Sold: 4398

   Item_ID  Units Sold
0     7827         981
1     3229         929
2     1473         909
3       79         661
4     5081         486
_______________________________
-------------------------------
Brand Number 3: Jabra
Brand Units Sold: 2090

   Item_ID  Units Sold
0     1750        1284
1     4811         588
2     3049         153
3     2243          65
_______________________________
-------------------------------
Brand Number 4: ebasy
Brand Units Sold: 1098

   Item_ID  Units Sold
0     4821         461
1     3576         256
2     1274         217
3     4036          95
4     7163          69
_______________________________
-------------------------------
Brand Number 5: Sennheiser
Brand Units Sold: 745

   Item_ID  Units Sold
0     1560         569
1     4590         122
2      227          54
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Television & Video, best selling brands recommended are:

Brand Number 1: Samsung
Brand Units Sold: 4022

   Item_ID  Units Sold
0     2129        4022
_______________________________
-------------------------------
Brand Number 2: Panasonic
Brand Units Sold: 1074

   Item_ID  Units Sold
0     1618         287
1     1296         228
2      476         189
3     1604         129
4     1563         109
_______________________________
-------------------------------
Brand Number 3: Toshiba
Brand Units Sold: 790

   Item_ID  Units Sold
0     4177         497
1     2173         155
2     2984          72
3     2214          66
_______________________________
-------------------------------
Brand Number 4: DBPOWER
Brand Units Sold: 456

   Item_ID  Units Sold
0     8074         296
1     4633         108
2     2654          52
_______________________________
-------------------------------
Brand Number 5: Bose
Brand Units Sold: 441

   Item_ID  Units Sold
0     5202         441
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Car Electronics & GPS, best selling brands recommended are:

Brand Number 1: Garmin
Brand Units Sold: 7054

   Item_ID  Units Sold
0      283        2863
1     1105        1457
2     4152         676
3     1323         421
4     7936         319
_______________________________
-------------------------------
Brand Number 2: Pyle
Brand Units Sold: 1259

   Item_ID  Units Sold
0      638         587
1     8041         126
2     8335         108
3     6858         105
4      593         103
_______________________________
-------------------------------
Brand Number 3: Uniden
Brand Units Sold: 638

   Item_ID  Units Sold
0     1056         567
1     6731          71
_______________________________
-------------------------------
Brand Number 4: Mpow
Brand Units Sold: 504

   Item_ID  Units Sold
0     5749         504
_______________________________
-------------------------------
Brand Number 5: JVC
Brand Units Sold: 220

   Item_ID  Units Sold
0     1037          77
1      980          73
2      441          70
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Home Audio, best selling brands recommended are:

Brand Number 1: TaoTronics
Brand Units Sold: 6216

   Item_ID  Units Sold
0     5054        4033
1     8375        1252
2     6328         931
_______________________________
-------------------------------
Brand Number 2: Pyle
Brand Units Sold: 3558

   Item_ID  Units Sold
0      827         743
1     1236         618
2     1125         600
3     3090         433
4     1013         432
_______________________________
-------------------------------
Brand Number 3: Sony
Brand Units Sold: 554

   Item_ID  Units Sold
0      511         554
_______________________________
-------------------------------
Brand Number 4: Bose
Brand Units Sold: 431

   Item_ID  Units Sold
0     3625         192
1      211         161
2      178          78
_______________________________
-------------------------------
Brand Number 5: EldHus
Brand Units Sold: 204

   Item_ID  Units Sold
0     6662         204
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Security & Surveillance, best selling brands recommended are:

Brand Number 1: Uniden
Brand Units Sold: 1018

   Item_ID  Units Sold
0     3989         529
1     1250         276
2     3603         111
3     6638         102
_______________________________
-------------------------------
Brand Number 2: Funlux
Brand Units Sold: 81

   Item_ID  Units Sold
0     4831          81
_______________________________
-------------------------------
Brand Number 3: Samsung
Brand Units Sold: 57

   Item_ID  Units Sold
0     3293          57
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Wearable Technology, best selling brands recommended are:

Brand Number 1: Sony
Brand Units Sold: 637

   Item_ID  Units Sold
0     5402         637
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Dresses, best selling brands recommended are:

Brand Number 1: Chi Chi London
Brand Units Sold: 2851

   Item_ID  Units Sold
0   107539         975
1    61110         369
2   126114         331
3   126098         171
4   135364         153
_______________________________
-------------------------------
Brand Number 2: Retrolicious
Brand Units Sold: 1904

   Item_ID  Units Sold
0    56780         621
1    28252         616
2    35525         310
3   119314         182
4    40141         105
_______________________________
-------------------------------
Brand Number 3: ModCloth
Brand Units Sold: 1797

   Item_ID  Units Sold
0    70538        1045
1   142002         120
2   137270         106
3   154064          85
4   150866          71
_______________________________
-------------------------------
Brand Number 4: Ryu
Brand Units Sold: 576

   Item_ID  Units Sold
0    48629         352
1   104356         143
2   110169          81
_______________________________
-------------------------------
Brand Number 5: Emily and Fin
Brand Units Sold: 384

   Item_ID  Units Sold
0   128563         210
1   153665          92
2   141295          82
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Outerwear, best selling brands recommended are:

Brand Number 1: ModCloth
Brand Units Sold: 1381

   Item_ID  Units Sold
0    58990         418
1    56289         208
2    78510         193
3    85507         192
4   123500         147
_______________________________
-------------------------------
Brand Number 2: Steve Madden
Brand Units Sold: 1172

   Item_ID  Units Sold
0    27439         772
1   110820         195
2   110824         116
3   129113          89
_______________________________
-------------------------------
Brand Number 3: Jack by BB Dakota
Brand Units Sold: 330

   Item_ID  Units Sold
0    71607         330
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Bottoms, best selling brands recommended are:

Brand Number 1: ModCloth
Brand Units Sold: 4111

   Item_ID  Units Sold
0    21296        1636
1    54222         749
2   121554         214
3   141880         168
4   153501         139
_______________________________
-------------------------------
Brand Number 2: Miss Candyfloss
Brand Units Sold: 307

   Item_ID  Units Sold
0   121272         148
1   133399         100
2   133539          59
_______________________________
-------------------------------
Brand Number 3: Pink Martini
Brand Units Sold: 68

   Item_ID  Units Sold
0   111781          68
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Tops, best selling brands recommended are:

Brand Number 1: ModCloth
Brand Units Sold: 3271

   Item_ID  Units Sold
0    40899         957
1    59176         933
2   118020         382
3   118948         169
4   112771         168
_______________________________
-------------------------------
Brand Number 2: Out of Print
Brand Units Sold: 696

   Item_ID  Units Sold
0    67022         448
1   113643         248
_______________________________
-------------------------------
Brand Number 3: Kin Ship
Brand Units Sold: 360

   Item_ID  Units Sold
0   122266         192
1    70230         168
_______________________________
-------------------------------
Brand Number 4: Banned
Brand Units Sold: 93

   Item_ID  Units Sold
0   146244          93
_______________________________
-------------------------------
Brand Number 5: Daisey Natives
Brand Units Sold: 58

   Item_ID  Units Sold
0   143764          58
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

# def bestselling_brand_items_ovr_1(category_name, brand_name):
#     """this function prints the dataframe showing items and the respective units sold for those items for a given category"""
#     temp_df = best_selling_item[
#         (best_selling_item.brand == brand_name)
#         & (best_selling_item.category == category_name)
#     ].head()
#     item_list = temp_df.item_id.tolist()
#     return item_list

# bestselling_brand_recommendation = pd.DataFrame(
#     columns=[
#         "Recommendation_System_Type",
#         "Category",
#         "Brand",
#         "Brand_Sales",
#         "Top_Products",
#     ]
# )

# for k in range(len(list_category)):
# #     print(f"For Category: {list_category[k]}, best selling brands recommended are:\n")
#     sol = best_selling_brand_ovr(list_category[k])
#     for j in range(len(sol[0])):
#         brand_id = sol[0][j]
#         brand_rating = sol[1][j]
# #         print(f"Brand Number {j+1}: {brand_id}")
# #         print(f"Brand Rating: {brand_rating}")
#         item_list_1 = bestselling_brand_items_ovr_1(list_category[k], brand_id)
        
#         bestselling_brand_recommendation.loc[len(bestselling_brand_recommendation)] = [
#             "Overall_Bestselling_Brand",
#             list_category[k],
#             brand_id,
#             brand_rating,
#             item_list_1
#         ]
        
# # bestselling_brand_recommendation

'Below code is to create the website for demonstration of Recommendation System, Hence it is commented out'

Now let us focus on the latest trend i.e. from the past 2 years

"""Forming the dataframe of past 2 years"""
best_selling_2years = best_selling_db[best_selling_db.timestamp > "2017-01-01"]

"""itemwise"""
best_selling_item_2years = (
    best_selling_2years.groupby(by=["item_id", "category", "brand"], as_index=False)
    .agg({"count": pd.Series.sum})
    .sort_values(by="count", ascending=False)
)

def top5_bestselling_items_2years(category_name):
    """this function prints dataframe showing items, their respective brands & respective units sold for given category"""
    df = best_selling_item_2years[best_selling_item.category == category_name].head()
    item_list = df.item_id.tolist()
    brand_list = df.brand.tolist()
    sold_list = df["count"].tolist()
    temp_df = pd.DataFrame(
        {"Item_ID": item_list, "Brand": brand_list, "Sold_Units": sold_list}
    )
    print(temp_df)

"""brandwise"""

"""For suggesting brand, we cannot suggest "Other" as brand as we have substitued "Other" for Missing values.
Hence, we will drop "Other" brand and consider the rest"""
best_selling_brand_raw_2years = best_selling_2years[best_selling_db.brand != "Other"]


"""Forming DataFrame"""

best_selling_brand_2years = (
    best_selling_brand_raw_2years.groupby(by=["category", "brand"], as_index=False)
    .agg({"count": pd.Series.sum})
    .sort_values(by="count", ascending=False)
)

<ipython-input-78-afde63db66b6>:5: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

def best_selling_brand_2years_fun(category_name):
    """this function returns list of best selling brands and the units sold for a given category"""
    temp_df = best_selling_brand_2years[
        best_selling_brand_2years.category == category_name
    ].head()
    brand_list = temp_df.brand.tolist()
    sold_list = temp_df["count"].tolist()
    return [brand_list, sold_list]


def bestselling_brand_items_2years(category_name, brand_name):
    """this function prints dataframe showing best-selling items & units sold for given category and brand"""
    temp_df = best_selling_item_2years[
        (best_selling_item_2years.brand == brand_name)
        & (best_selling_item_2years.category == category_name)
    ].head()
    item_list = temp_df.item_id.tolist()
    sold_list = temp_df["count"].tolist()
    df = pd.DataFrame({"Item_ID": item_list, "Units Sold": sold_list})
    print(df)

5.c. Latest Trend: Best Selling Items from 2017 to 2019

print("Recent Trends: Best Selling Items from 2017 to 2019:")
print("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n\n")
for i in range(len(list_category)):
    print(f"For Category:{list_category[i]}")
    print()
    top5_bestselling_items_2years(list_category[i])
    print("_______________________________")
    print()
    print()

Recent Trends: Best Selling Items from 2017 to 2019:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


For Category:Portable Audio & Video

   Item_ID     Brand  Sold_Units
0     2031  Logitech        3042
1     7022      Bose        2043
2     2237     Other        1881
3     5897     Other        1735
4      647  Logitech        1170
_______________________________


For Category:Computers & Accessories

   Item_ID    Brand  Sold_Units
0     6819     Mpow        3734
1     5063    Other        3312
2     4009  DBPOWER        2772
3     6689     Mpow        1592
4     5041     Bose        1190
_______________________________


For Category:Headphones

   Item_ID  Brand  Sold_Units
0     6378  Other        3234
1     6210  Other        2873
2     5795  Other        2281
3     5884  Other        1856
4     6818  Other        1818
_______________________________


For Category:Camera & Photo

   Item_ID       Brand  Sold_Units
0      162  Sennheiser        2575
1     5054  TaoTronics        1388
2      274       Other        1346
3     5301       Other        1151
4     6535       Other        1106
_______________________________


For Category:Accessories & Supplies

   Item_ID  Brand  Sold_Units
0     7780  Other       15639
1     1034  Other        1319
2     1706  Other        1064
3     3106   Bose         980
4     5548  Other         919
_______________________________


For Category:Television & Video

   Item_ID  Brand  Sold_Units
0     8802  Other         489
1     7730  Other         449
2     1575   Sony         374
3     6036  Other         330
4     7291  Other         211
_______________________________


For Category:Car Electronics & GPS

   Item_ID  Brand  Sold_Units
0     7770  Other        1033
1     3930  Other         506
2     3649  Other         414
3     8851  Other         394
4     7107  Other         384
_______________________________


For Category:Home Audio

   Item_ID    Brand  Sold_Units
0     7206    Other        2068
1     2793    Other         382
2     2837    Other         257
3     6396  Generic         187
4     1211    Other         182
_______________________________


For Category:Security & Surveillance

   Item_ID  Brand  Sold_Units
0     6170  Other        3214
1     3201  Other         363
2     7055  Other         279
3     2678  Other         266
4     6744  Other         235
_______________________________


For Category:Wearable Technology

   Item_ID  Brand  Sold_Units
0     4718  Other        1665
1     4969   Bose         812
2     5914  Other         196
3     4834  Other         170
4     5808  Other         166
_______________________________


For Category:Dresses

   Item_ID  Brand  Sold_Units
0     8436  Other          27
1     7341  Other           5
_______________________________


For Category:Outerwear

Empty DataFrame
Columns: [Item_ID, Brand, Sold_Units]
Index: []
_______________________________


For Category:Bottoms

Empty DataFrame
Columns: [Item_ID, Brand, Sold_Units]
Index: []
_______________________________


For Category:Tops

Empty DataFrame
Columns: [Item_ID, Brand, Sold_Units]
Index: []
_______________________________

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

<ipython-input-77-cb13e6549788>:3: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

# def top5_bestselling_items_2years_1(category_name):
#     """this function prints dataframe showing items, their respective brands & respective units sold for given category"""
#     df = best_selling_item_2years[best_selling_item.category == category_name].head()
#     item_list = df.item_id.tolist()
#     brand_list = df.brand.tolist()
#     sold_list = df["count"].tolist()
#     temp_df = pd.DataFrame(
#         {"Item_ID": item_list, "Brand": brand_list, "Sold_Units": sold_list}
#     )
#     return temp_df

# temp_df3 = pd.DataFrame(columns=["Category", "Item_ID", "Brand", "Sold_Units"])


# for i in range(len(list_category)):
# #     print(f"For Category:{list_category[i]}")
#     temp_df = top5_bestselling_items_over_1(list_category[i])
#     temp_df.insert(0,"Category",list_category[i])
#     temp_df3 = temp_df3.append(temp_df)
    
# temp_df3.insert(0,"Recommendation_Type","Latest_Bestselling_Items")

# bestselling_item_recommendation = bestselling_item_recommendation.append(temp_df3.copy())

# bestselling_item_recommendation.to_csv("bestselling_item_recommendation.csv")

'Below code is to create the website for demonstration of Recommendation System, Hence it is commented out'

5.d. Latest Trend: Brandwise Best Selling Items from 2017 to 2019

print("Recommending Items as per Top Selling Brand from 2017 to 2019:")
print("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n\n")

for k in range(len(list_category)):
    print(f"For Category: {list_category[k]}, best selling brands are:\n")
    sol = best_selling_brand_2years_fun(list_category[k])
    for j in range(len(sol[0])):
        brand_id = sol[0][j]
        brand_rating = sol[1][j]
        print(f"Brand Rank {j+1}: {brand_id}")
        print(f"Brand Units Sold: {brand_rating}")
        print()
        bestselling_brand_items_2years(list_category[k], brand_id)
        print("_______________________________")
        print("-------------------------------")
    print()
    print(
        "________________________________________________________________________________________________"
    )
    print(
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
    )
    print()

Recommending Items as per Top Selling Brand from 2017 to 2019:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


For Category: Portable Audio & Video, best selling brands are:

Brand Rank 1: DBPOWER
Brand Units Sold: 3033

   Item_ID  Units Sold
0     4009        2772
1     2864         258
2     4079           3
_______________________________
-------------------------------
Brand Rank 2: Bose
Brand Units Sold: 2527

   Item_ID  Units Sold
0     7022        2043
1     5171         134
2     3195         130
3     5273         106
4     5156          76
_______________________________
-------------------------------
Brand Rank 3: Pyle
Brand Units Sold: 677

   Item_ID  Units Sold
0     3468         148
1     4753         131
2     5445          97
3       33          56
4     5740          49
_______________________________
-------------------------------
Brand Rank 4: Uniden
Brand Units Sold: 606

   Item_ID  Units Sold
0     2223         210
1     2159         170
2     4419         110
3     2434          95
4     2298           6
_______________________________
-------------------------------
Brand Rank 5: Mpow
Brand Units Sold: 432

   Item_ID  Units Sold
0     5748         429
1     5974           3
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Computers & Accessories, best selling brands are:

Brand Rank 1: Logitech
Brand Units Sold: 6269

   Item_ID  Units Sold
0     2031        3042
1      647        1170
2     1124         635
3     1780         452
4     3630         241
_______________________________
-------------------------------
Brand Rank 2: Fintie
Brand Units Sold: 1161

   Item_ID  Units Sold
0     8736         585
1     3643         225
2     5200          66
3     7901          57
4     5354          51
_______________________________
-------------------------------
Brand Rank 3: Bose
Brand Units Sold: 980

   Item_ID  Units Sold
0     3106         980
_______________________________
-------------------------------
Brand Rank 4: LSS
Brand Units Sold: 424

   Item_ID  Units Sold
0     3884         184
1     4543          64
2     4271          44
3     3951          43
4     4043          33
_______________________________
-------------------------------
Brand Rank 5: Apple
Brand Units Sold: 416

   Item_ID  Units Sold
0     5364         104
1     5329          83
2     6562          80
3     5405          65
4     4302          49
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Headphones, best selling brands are:

Brand Rank 1: Mpow
Brand Units Sold: 5832

   Item_ID  Units Sold
0     6819        3734
1     6689        1592
2     6835         158
3     7809         150
4     5424          43
_______________________________
-------------------------------
Brand Rank 2: Skullcandy
Brand Units Sold: 4181

   Item_ID  Units Sold
0     5178        1107
1     2355         989
2     8195         802
3     5261         287
4     8350         282
_______________________________
-------------------------------
Brand Rank 3: Bose
Brand Units Sold: 4108

   Item_ID  Units Sold
0     5041        1190
1     7828         829
2     4969         812
3     7869         659
4     7645         253
_______________________________
-------------------------------
Brand Rank 4: Sennheiser
Brand Units Sold: 3238

   Item_ID  Units Sold
0      162        2575
1      259         201
2     1242         105
3     2228          72
4     2032          72
_______________________________
-------------------------------
Brand Rank 5: Panasonic
Brand Units Sold: 1411

   Item_ID  Units Sold
0      111         932
1     9510         233
2     9380         107
3     9537         105
4     3248          34
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Camera & Photo, best selling brands are:

Brand Rank 1: Neewer
Brand Units Sold: 2160

   Item_ID  Units Sold
0     8341         920
1     1605         522
2     8077         379
3     3783          70
4     7101          57
_______________________________
-------------------------------
Brand Rank 2: Fujifilm
Brand Units Sold: 1923

   Item_ID  Units Sold
0     1258         342
1     7349         198
2     7895         182
3     6444         116
4     3384         110
_______________________________
-------------------------------
Brand Rank 3: Polaroid
Brand Units Sold: 1774

   Item_ID  Units Sold
0     6726         840
1     2586         768
2     2366          92
3     1327          48
4     1531          12
_______________________________
-------------------------------
Brand Rank 4: Nikon
Brand Units Sold: 738

   Item_ID  Units Sold
0     7544         225
1     6745         216
2     8609          58
3     8539          55
4     6599          40
_______________________________
-------------------------------
Brand Rank 5: Canon
Brand Units Sold: 447

   Item_ID  Units Sold
0     9225         141
1     9189          73
2     8484          59
3      220          43
4     6434          27
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Accessories & Supplies, best selling brands are:

Brand Rank 1: EldHus
Brand Units Sold: 989

   Item_ID  Units Sold
0     8282         358
1     2138         355
2     9188          70
3     1861          46
4     9169          36
_______________________________
-------------------------------
Brand Rank 2: Etre Jeune
Brand Units Sold: 276

   Item_ID  Units Sold
0     7827         103
1     5081          75
2     9244          27
3     3229          24
4     1473          23
_______________________________
-------------------------------
Brand Rank 3: Jabra
Brand Units Sold: 245

   Item_ID  Units Sold
0     4811         187
1     3049          43
2     2243          11
3     1750           4
_______________________________
-------------------------------
Brand Rank 4: Sennheiser
Brand Units Sold: 101

   Item_ID  Units Sold
0     1560          68
1     4590          33
_______________________________
-------------------------------
Brand Rank 5: Pyle
Brand Units Sold: 79

   Item_ID  Units Sold
0     4225          64
1     1744           9
2     5258           6
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Television & Video, best selling brands are:

Brand Rank 1: Bose
Brand Units Sold: 201

   Item_ID  Units Sold
0     5202         201
_______________________________
-------------------------------
Brand Rank 2: DBPOWER
Brand Units Sold: 88

   Item_ID  Units Sold
0     8074          66
1     4633          20
2     2654           2
_______________________________
-------------------------------
Brand Rank 3: Samsung
Brand Units Sold: 63

   Item_ID  Units Sold
0     2129          63
_______________________________
-------------------------------
Brand Rank 4: Pyle
Brand Units Sold: 40

   Item_ID  Units Sold
0     1984          40
_______________________________
-------------------------------
Brand Rank 5: Toshiba
Brand Units Sold: 14

   Item_ID  Units Sold
0     4177          12
1     2173           1
2     2214           1
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Car Electronics & GPS, best selling brands are:

Brand Rank 1: Garmin
Brand Units Sold: 537

   Item_ID  Units Sold
0     7936         186
1     5185          83
2     4152          83
3     1105          75
4     6625          58
_______________________________
-------------------------------
Brand Rank 2: Pyle
Brand Units Sold: 286

   Item_ID  Units Sold
0      638          86
1     8041          72
2     8335          60
3     2431          24
4      593          21
_______________________________
-------------------------------
Brand Rank 3: Uniden
Brand Units Sold: 163

   Item_ID  Units Sold
0     1056         143
1     6731          20
_______________________________
-------------------------------
Brand Rank 4: Mpow
Brand Units Sold: 24

   Item_ID  Units Sold
0     5749          24
_______________________________
-------------------------------
Brand Rank 5: EINCAR
Brand Units Sold: 10

   Item_ID  Units Sold
0     6750          10
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Home Audio, best selling brands are:

Brand Rank 1: TaoTronics
Brand Units Sold: 2541

   Item_ID  Units Sold
0     5054        1388
1     8375        1099
2     6328          54
_______________________________
-------------------------------
Brand Rank 2: Pyle
Brand Units Sold: 595

   Item_ID  Units Sold
0     1236         116
1     3090         114
2     1125          86
3      827          79
4     5459          77
_______________________________
-------------------------------
Brand Rank 3: Sony
Brand Units Sold: 28

   Item_ID  Units Sold
0      511          28
_______________________________
-------------------------------
Brand Rank 4: Bose
Brand Units Sold: 16

   Item_ID  Units Sold
0     3625           9
1      211           5
2      178           2
_______________________________
-------------------------------
Brand Rank 5: EldHus
Brand Units Sold: 7

   Item_ID  Units Sold
0     6662           7
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Security & Surveillance, best selling brands are:

Brand Rank 1: Uniden
Brand Units Sold: 243

   Item_ID  Units Sold
0     3989         175
1     6638          52
2     1250          12
3     3603           4
_______________________________
-------------------------------
Brand Rank 2: Samsung
Brand Units Sold: 12

   Item_ID  Units Sold
0     3293          12
_______________________________
-------------------------------
Brand Rank 3: Funlux
Brand Units Sold: 7

   Item_ID  Units Sold
0     4831           7
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Wearable Technology, best selling brands are:

Brand Rank 1: Sony
Brand Units Sold: 123

   Item_ID  Units Sold
0     5402         123
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Dresses, best selling brands are:

Brand Rank 1: Chi Chi London
Brand Units Sold: 1084

   Item_ID  Units Sold
0   107539         286
1   126114         148
2   135364         107
3   141856         104
4   147244          86
_______________________________
-------------------------------
Brand Rank 2: ModCloth
Brand Units Sold: 848

   Item_ID  Units Sold
0    70538         187
1   142002         120
2   154064          85
3   150866          71
4   153445          67
_______________________________
-------------------------------
Brand Rank 3: Retrolicious
Brand Units Sold: 327

   Item_ID  Units Sold
0    40141         105
1   119314          79
2    56780          51
3    28252          34
4   140020          34
_______________________________
-------------------------------
Brand Rank 4: Emily and Fin
Brand Units Sold: 271

   Item_ID  Units Sold
0   128563          97
1   153665          92
2   141295          82
_______________________________
-------------------------------
Brand Rank 5: Pepaloves
Brand Units Sold: 183

   Item_ID  Units Sold
0   153481         183
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Outerwear, best selling brands are:

Brand Rank 1: ModCloth
Brand Units Sold: 384

   Item_ID  Units Sold
0   135555          73
1    86074          53
2   123500          53
3    58990          49
4    85507          44
_______________________________
-------------------------------
Brand Rank 2: Steve Madden
Brand Units Sold: 239

   Item_ID  Units Sold
0    27439         126
1   110820          45
2   129113          44
3   110824          24
_______________________________
-------------------------------
Brand Rank 3: Jack by BB Dakota
Brand Units Sold: 85

   Item_ID  Units Sold
0    71607          85
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Bottoms, best selling brands are:

Brand Rank 1: ModCloth
Brand Units Sold: 1642

   Item_ID  Units Sold
0    21296         150
1   153501         139
2   141880         136
3    54222         116
4   148257         107
_______________________________
-------------------------------
Brand Rank 2: Miss Candyfloss
Brand Units Sold: 153

   Item_ID  Units Sold
0   133399          68
1   121272          56
2   133539          29
_______________________________
-------------------------------
Brand Rank 3: Pink Martini
Brand Units Sold: 26

   Item_ID  Units Sold
0   111781          26
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

For Category: Tops, best selling brands are:

Brand Rank 1: ModCloth
Brand Units Sold: 933

   Item_ID  Units Sold
0    59176         142
1    40899         141
2   118020         108
3   153726          98
4   153517          87
_______________________________
-------------------------------
Brand Rank 2: Out of Print
Brand Units Sold: 52

   Item_ID  Units Sold
0    67022          37
1   113643          15
_______________________________
-------------------------------
Brand Rank 3: Kin Ship
Brand Units Sold: 51

   Item_ID  Units Sold
0   122266          45
1    70230           6
_______________________________
-------------------------------
Brand Rank 4: Banned
Brand Units Sold: 48

   Item_ID  Units Sold
0   146244          48
_______________________________
-------------------------------
Brand Rank 5: Daisey Natives
Brand Units Sold: 38

   Item_ID  Units Sold
0   143764          38
_______________________________
-------------------------------

________________________________________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

# # bestselling_brand_recommendation
# def bestselling_brand_items_2years_1(category_name, brand_name):
#     """this function prints dataframe showing best-selling items & units sold for given category and brand"""
#     temp_df = best_selling_item_2years[
#         (best_selling_item_2years.brand == brand_name)
#         & (best_selling_item_2years.category == category_name)
#     ].head()
#     item_list = temp_df.item_id.tolist()
#     return item_list


# temp_df = pd.DataFrame(
#     columns=[
#         "Recommendation_System_Type",
#         "Category",
#         "Brand",
#         "Brand_Sales",
#         "Top_Products",
#     ]
# )

# for k in range(len(list_category)):
# #     print(f"For Category: {list_category[k]}, best selling brands are:\n")
#     sol = best_selling_brand_2years_fun(list_category[k])
#     for j in range(len(sol[0])):
#         brand_id = sol[0][j]
#         brand_rating = sol[1][j]
# #         print(f"Brand Rank {j+1}: {brand_id}")
# #         print(f"Brand Rating: {brand_rating}")
#         temp_df.loc[len(temp_df)] = [
#             "Latest_Bestselling_Brand",
#             list_category[k],
#             brand_id,
#             brand_rating,
#             bestselling_brand_items_2years_1(list_category[k], brand_id),
#         ]
        
# bestselling_brand_recommendation = bestselling_brand_recommendation.append(temp_df)

# bestselling_brand_recommendation.to_csv("bestselling_brand_recommendation.csv")

'Below code is to create the website for demonstration of Recommendation System, Hence it is commented out'

6. Applying Market Basket Analysis (using Apriori algorithm)

for market basket, we will consider only those cases where the customers have purchased more than one item.

This will allow us to link the preferences/likenesses of user to buy an item based on the item purchased before.

For Market Basket Analysis / Apriori, higher the lift value, higher is the chance that the user will buy the consequent item based on the antecedant item, i.e. you can recommend the consequent item bought by the user upon the user selecting the antecedant item.

"""Items bought per category"""
consolidated_db.groupby(by="category", as_index=False).agg({"count": pd.Series.sum})

	category	count
0	Accessories & Supplies	158598
1	Bottoms	23625
2	Camera & Photo	192573
3	Car Electronics & GPS	33070
4	Computers & Accessories	322938
5	Dresses	34160
6	Headphones	359334
7	Home Audio	24193
8	Outerwear	7131
9	Portable Audio & Video	143370
10	Security & Surveillance	10806
11	Television & Video	32057
12	Tops	34977
13	Wearable Technology	16015

"""We have to convert the item_id and user_id to string"""
consolidated_db.item_id = consolidated_db.item_id.astype("str")
consolidated_db.user_id = consolidated_db.user_id.astype("str")

"""through trial and error, the max value of minimum-support-values which gives association rules are as below:"""

cat_support = {
    "Portable Audio & Video": 0.02,
    "Computers & Accessories": 0.005,
    "Headphones": 0.003,
    "Camera & Photo": 0.02,
    "Accessories & Supplies": 0.05,
    "Television & Video": 0.02,
    "Car Electronics & GPS": 0.02,
    "Home Audio": 0.05,
    "Security & Surveillance": 0.08,
    "Wearable Technology": 0.05,
    "Dresses": 0.046,
    "Outerwear": 0.1,
    "Bottoms": 0.05,
    "Tops": 0.04,
}

6.a. Item Recommendation per Category based on Market Basket Analysis (Apriori)

print("Categorical Item Suggestion based on Market Basket Analysis is:")
print()
for category_name in list_category:
    consolidated_db_temp = consolidated_db[consolidated_db.category == category_name]
    user_transactions_temp = consolidated_db_temp.groupby(
        by="user_id", as_index=False
    ).agg({"count": pd.Series.sum})

    user_transactions_temp.columns = ["user_id", "number_of_transactions"]
    temp_basket = pd.merge(consolidated_db_temp, user_transactions_temp, on="user_id")
    temp_basket = temp_basket[temp_basket.number_of_transactions > 1]

    temp_basket["concat_item"] = temp_basket.item_id + "|"
    temp_basket_trans = temp_basket.groupby(by="user_id", as_index=False).agg(
        {"concat_item": pd.Series.sum}
    )

    temp_basket_trans.concat_item = [
        i.split("|") for i in temp_basket_trans.concat_item.tolist()
    ]
    transactions_list = temp_basket_trans.concat_item.tolist()
    transactions_list = [
        transactions_list[i][0 : (len(transactions_list[i]) - 1)]
        for i in range(len(transactions_list))
    ]

    # We form an array where every column has one separate column with purchase data in rows as boolean
    encode = TransactionEncoder()
    array_transactions = encode.fit_transform(transactions_list)
    # items as columns can be seen via
    item_cols = encode.columns_

    # We form a dataframe
    market_basket_clean = pd.DataFrame(array_transactions, columns=item_cols)

    # We apply apriori algorithm to see the connection between items
    model = apriori(
        market_basket_clean,
        min_support=cat_support[category_name],
        max_len=5,
        use_colnames=True,
    )
    rule = association_rules(model, metric="lift", min_threshold=1)
    print(f"For category: {category_name}")
    print()
    print(
        rule.sort_values(by="lift", ascending=False)[
            ["antecedents", "consequents", "support", "confidence", "lift"]
        ]
    )
    print("____________________________________________________________________")
    print("XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX")
    print()
    print()

Categorical Item Suggestion based on Market Basket Analysis is:

For category: Portable Audio & Video

  antecedents consequents   support  confidence       lift
0      (1706)      (3035)  0.023612    0.477876  15.612895
1      (3035)      (1706)  0.023612    0.771429  15.612895
3      (7780)      (8711)  0.030171    0.123656   3.008522
2      (8711)      (7780)  0.030171    0.734043   3.008522
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Computers & Accessories

  antecedents consequents   support  confidence       lift
0      (2523)      (2486)  0.007543    0.611842  16.876987
1      (2486)      (2523)  0.007543    0.208054  16.876987
3      (4718)       (274)  0.006732    0.188209   3.683511
2       (274)      (4718)  0.006732    0.131746   3.683511
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Headphones

  antecedents consequents   support  confidence       lift
4      (9315)      (7256)  0.003426    0.441441  21.118322
5      (7256)      (9315)  0.003426    0.163880  21.118322
2       (111)      (6210)  0.003216    0.116751   3.976215
3      (6210)       (111)  0.003216    0.109524   3.976215
0       (111)      (1575)  0.003985    0.144670   3.477917
1      (1575)       (111)  0.003985    0.095798   3.477917
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Camera & Photo

  antecedents consequents   support  confidence       lift
0      (6726)      (2586)  0.037314    0.836830  16.006242
1      (2586)      (6726)  0.037314    0.713718  16.006242
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Accessories & Supplies

  antecedents consequents   support  confidence      lift
4      (3995)      (4806)  0.090031    0.559157  3.043594
5      (4806)      (3995)  0.090031    0.490057  3.043594
2      (4806)      (2483)  0.110647    0.602273  2.989520
3      (2483)      (4806)  0.110647    0.549223  2.989520
1      (2483)      (3995)  0.072286    0.358808  2.228450
0      (3995)      (2483)  0.072286    0.448947  2.228450
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Television & Video

  antecedents consequents   support  confidence       lift
2      (6078)      (9144)  0.021552    0.833333  38.666667
3      (9144)      (6078)  0.021552    1.000000  38.666667
0      (3617)      (2129)  0.030172    0.368421   2.947368
1      (2129)      (3617)  0.030172    0.241379   2.947368
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Car Electronics & GPS

  antecedents consequents   support  confidence       lift
2       (745)      (1115)  0.048780    0.769231  15.769231
3      (1115)       (745)  0.048780    1.000000  15.769231
0      (4152)      (1105)  0.034146    0.466667   5.314815
1      (1105)      (4152)  0.034146    0.388889   5.314815
4      (6005)      (3728)  0.024390    0.200000   1.640000
5      (3728)      (6005)  0.024390    0.200000   1.640000
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Home Audio

  antecedents consequents   support  confidence      lift
0      (5054)      (6328)  0.062016    0.170213  1.829787
1      (6328)      (5054)  0.062016    0.666667  1.829787
2      (5054)      (8802)  0.054264    0.148936  1.200798
3      (8802)      (5054)  0.054264    0.437500  1.200798
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Security & Surveillance

  antecedents consequents   support  confidence      lift
2      (5283)      (6156)  0.083333    0.500000  3.818182
3      (6156)      (5283)  0.083333    0.636364  3.818182
0      (5283)      (5456)  0.095238    0.571429  3.000000
1      (5456)      (5283)  0.095238    0.500000  3.000000
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Wearable Technology

  antecedents consequents  support  confidence      lift
0      (4278)      (6003)  0.08046    0.538462  3.346154
1      (6003)      (4278)  0.08046    0.500000  3.346154
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Dresses

  antecedents consequents   support  confidence      lift
0     (32406)     (16411)  0.046748    0.315068  2.226409
1     (16411)     (32406)  0.046748    0.330341  2.226409
2     (32405)     (32406)  0.047764    0.306688  2.066996
3     (32406)     (32405)  0.047764    0.321918  2.066996
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Outerwear

  antecedents consequents   support  confidence      lift
4     (27439)     (32236)  0.112583    0.431472  1.428778
5     (32236)     (27439)  0.112583    0.372807  1.428778
2     (32236)     (11960)  0.156291    0.517544  1.400522
3     (11960)     (32236)  0.156291    0.422939  1.400522
0     (27439)     (11960)  0.119205    0.456853  1.236286
1     (11960)     (27439)  0.119205    0.322581  1.236286
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Bottoms

  antecedents consequents   support  confidence      lift
6     (80238)     (62447)  0.055031    0.366876  2.466526
7     (62447)     (80238)  0.055031    0.369979  2.466526
2     (21296)     (54222)  0.055346    0.243430  2.058798
3     (54222)     (21296)  0.055346    0.468085  2.058798
0     (21296)     (50818)  0.057862    0.254495  1.675558
1     (50818)     (21296)  0.057862    0.380952  1.675558
4     (21296)     (62447)  0.052516    0.230982  1.552902
5     (62447)     (21296)  0.052516    0.353066  1.552902
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


For category: Tops

  antecedents consequents   support  confidence      lift
0     (40899)     (34935)  0.041047    0.391304  2.185324
1     (34935)     (40899)  0.041047    0.229236  2.185324
2     (57369)     (34935)  0.040254    0.351820  1.964814
3     (34935)     (57369)  0.040254    0.224806  1.964814
____________________________________________________________________
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

# # print("Categorical Item Suggestion based on Market Basket Analysis is:")
# # print()
# apriori_item_df = pd.DataFrame(columns=["antecedents", "consequents", "support", "confidence", "lift", "Category"])
# for category_name in list_category:
#     consolidated_db_temp = consolidated_db[consolidated_db.category == category_name]
#     user_transactions_temp = consolidated_db_temp.groupby(
#         by="user_id", as_index=False
#     ).agg({"count": pd.Series.sum})

#     user_transactions_temp.columns = ["user_id", "number_of_transactions"]
#     temp_basket = pd.merge(consolidated_db_temp, user_transactions_temp, on="user_id")
#     temp_basket = temp_basket[temp_basket.number_of_transactions > 1]

#     temp_basket["concat_item"] = temp_basket.item_id + "|"
#     temp_basket_trans = temp_basket.groupby(by="user_id", as_index=False).agg(
#         {"concat_item": pd.Series.sum}
#     )

#     temp_basket_trans.concat_item = [
#         i.split("|") for i in temp_basket_trans.concat_item.tolist()
#     ]
#     transactions_list = temp_basket_trans.concat_item.tolist()
#     transactions_list = [
#         transactions_list[i][0 : (len(transactions_list[i]) - 1)]
#         for i in range(len(transactions_list))
#     ]

#     # We form an array where every column has one separate column with purchase data in rows as boolean
#     encode = TransactionEncoder()
#     array_transactions = encode.fit_transform(transactions_list)
#     # items as columns can be seen via
#     item_cols = encode.columns_

#     # We form a dataframe
#     market_basket_clean = pd.DataFrame(array_transactions, columns=item_cols)

#     # We apply apriori algorithm to see the connection between items
#     model = apriori(
#         market_basket_clean,
#         min_support=cat_support[category_name],
#         max_len=5,
#         use_colnames=True,
#     )
#     rule = association_rules(model, metric="lift", min_threshold=1)
# #     print(f"For category: {category_name}")
# #     print()
#     rule = rule.sort_values(by="lift", ascending=False)[
#             ["antecedents", "consequents", "support", "confidence", "lift"]
#         ]
# #     print("____________________________________________________________________")
# #     print("XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX")
# #     print()
# #     print()
#     rule["Category"] = category_name
#     apriori_item_df = apriori_item_df.append(rule)
    
# apriori_item_df.to_csv("apriori_item.csv")

'Below code is to create the website for demonstration of Recommendation System, Hence it is commented out'

---

Now that we have established relation between products, we will now check if relation exists between categories, i.e. if a person has bought item from category A, we will try to find the likelihood that he will buy item from category B

For this purpose, we cannot take consolidated dataframe as transactions for electric database is far greater than modcloth

We have to establish the basket analysis association rules separately for these 2 dataframes

We have to drop users who have made only 1 transaction as we cannot use these cases for market basket analysis

"""We will be considering dataframes: select_electronics_db & select_modcloth_db"""
select_electronics_db["count"] = 1
select_modcloth_db["count"] = 1


"""Forming a dataframe of electronics and modcloth transactions per user"""
electronics_temp = select_electronics_db.groupby(by="user_id", as_index=False).agg(
    {"count": pd.Series.sum}
)
modcloth_temp = select_modcloth_db.groupby(by="user_id", as_index=False).agg(
    {"count": pd.Series.sum}
)


"""We will rename the column 'count' to 'transactions' to avoid any confusion as we will be merging this dataframe
with parent databrame"""
electronics_temp.rename(columns={"count": "transactions"}, inplace=True)
modcloth_temp.rename(columns={"count": "transactions"}, inplace=True)


"""merging this dataframe with parent dataframe"""
select_electronics_db = pd.merge(select_electronics_db, electronics_temp, on="user_id")
select_modcloth_db = pd.merge(select_modcloth_db, modcloth_temp, on="user_id")

<ipython-input-89-a1ede762dd03>:2: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

<ipython-input-89-a1ede762dd03>:3: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

We will drop the users with only 1 transaction

"""dropping users with only 1 transaction"""
select_electronics_db = select_electronics_db[select_electronics_db.transactions > 1]
select_modcloth_db = select_modcloth_db[select_modcloth_db.transactions > 1]

select_electronics_db.head(2)

	item_id	user_id	rating	timestamp	category	brand	count	transactions
7	0	7	2.0	1999-07-13	Portable Audio & Video	Other	1	3
8	2205	7	5.0	2015-06-18	Computers & Accessories	Other	1	3

select_modcloth_db.head(2)

	item_id	user_id	rating	timestamp	category	brand	count	transactions
0	7443	Alex	4	2010-01-21 08:00:00+00:00	Dresses	Other	1	66
1	21296	Alex	5	2011-10-13 07:00:00+00:00	Bottoms	ModCloth	1	66

6.b. Category Recommendation based on Market Basket Analysis (Apriori)

Electronics Database

temp_db_cat = select_electronics_db.copy()
temp_db_cat["concat_cat"] = temp_db_cat.category + "|"
temp_db_cat = temp_db_cat.groupby(by="user_id", as_index=False).agg(
    {"concat_cat": pd.Series.sum}
)
temp_db_cat.concat_cat = [i.split("|") for i in temp_db_cat.concat_cat]

# to remove duplicates in the transaction, we use the 'set' type casting on the list before converting it back to list
temp_db_cat.concat_cat = [list(set(i[: len(i) - 1])) for i in temp_db_cat.concat_cat]

# this will have certain users where the user has purchases items from only one category and not from 2 or more categories
# for forming association rules, we have to consider cases where user has purchased items from 2 or more categories
temp_db_cat["cat_num"] = [len(i) for i in temp_db_cat.concat_cat]
temp_db_cat = temp_db_cat[temp_db_cat.cat_num > 1]

category_transaction_list = temp_db_cat.concat_cat.tolist()

encode_cat = TransactionEncoder()
array_transactions_cat = encode_cat.fit_transform(category_transaction_list)
# items as columns can be seen via
cat_cols = encode_cat.columns_

# We form a dataframe of categories
market_basket_cat_clean = pd.DataFrame(array_transactions_cat, columns=cat_cols)

# We apply apriori algorithm to see the connection between items
model_cat = apriori(
    market_basket_cat_clean, min_support=0.0013, max_len=5, use_colnames=True
)
rule_cat = association_rules(model_cat, metric="lift", min_threshold=1)
rule_cat.sort_values(by="lift", ascending=False)[
    ["antecedents", "consequents", "confidence", "lift"]
]

	antecedents	consequents	confidence	lift
13	(Headphones, Camera & Photo, Accessories & Sup...	(Computers & Accessories, Portable Audio & Video)	0.127848	1.601001
32	(Computers & Accessories, Portable Audio & Video)	(Headphones, Camera & Photo, Accessories & Sup...	0.017095	1.601001
28	(Portable Audio & Video, Accessories & Supplies)	(Headphones, Camera & Photo, Computers & Acces...	0.032053	1.586243
17	(Headphones, Camera & Photo, Computers & Acces...	(Portable Audio & Video, Accessories & Supplies)	0.067559	1.586243
21	(Computers & Accessories, Portable Audio & Vid...	(Headphones, Camera & Photo)	0.158307	1.541079
24	(Headphones, Camera & Photo)	(Computers & Accessories, Portable Audio & Vid...	0.013289	1.541079
36	(Portable Audio & Video)	(Headphones, Camera & Photo, Computers & Acces...	0.005709	1.441663
9	(Headphones, Camera & Photo, Computers & Acces...	(Portable Audio & Video)	0.344710	1.441663
15	(Headphones, Computers & Accessories, Accessor...	(Camera & Photo, Portable Audio & Video)	0.061736	1.416710
30	(Camera & Photo, Portable Audio & Video)	(Headphones, Computers & Accessories, Accessor...	0.031328	1.416710
18	(Headphones, Computers & Accessories, Portable...	(Camera & Photo, Accessories & Supplies)	0.082991	1.408581
27	(Camera & Photo, Accessories & Supplies)	(Headphones, Computers & Accessories, Portable...	0.023170	1.408581
19	(Camera & Photo, Portable Audio & Video, Acces...	(Headphones, Computers & Accessories)	0.299703	1.334371
26	(Headphones, Computers & Accessories)	(Camera & Photo, Portable Audio & Video, Acces...	0.006078	1.334371
10	(Headphones, Computers & Accessories, Portable...	(Camera & Photo)	0.384030	1.287597
35	(Camera & Photo)	(Headphones, Computers & Accessories, Portable...	0.004577	1.287597
1	(Accessories & Supplies)	(Wearable Technology)	0.042844	1.266386
0	(Wearable Technology)	(Accessories & Supplies)	0.407111	1.266386
22	(Camera & Photo, Portable Audio & Video, Compu...	(Headphones, Accessories & Supplies)	0.151198	1.259566
23	(Headphones, Accessories & Supplies)	(Camera & Photo, Portable Audio & Video, Compu...	0.011373	1.259566
31	(Camera & Photo, Computers & Accessories)	(Headphones, Portable Audio & Video, Accessori...	0.011210	1.254680
14	(Headphones, Portable Audio & Video, Accessori...	(Camera & Photo, Computers & Accessories)	0.152799	1.254680
16	(Headphones, Camera & Photo, Portable Audio & ...	(Computers & Accessories, Accessories & Supplies)	0.150522	1.207961
29	(Computers & Accessories, Accessories & Supplies)	(Headphones, Camera & Photo, Portable Audio & ...	0.010956	1.207961
8	(Headphones, Camera & Photo, Portable Audio & ...	(Computers & Accessories)	0.627329	1.198820
37	(Computers & Accessories)	(Headphones, Camera & Photo, Portable Audio & ...	0.002609	1.198820
12	(Camera & Photo, Portable Audio & Video, Compu...	(Headphones)	0.608434	1.165662
33	(Headphones)	(Camera & Photo, Portable Audio & Video, Compu...	0.002615	1.165662
7	(Computers & Accessories)	(Headphones, Camera & Photo, Television & Video)	0.002738	1.164167
2	(Headphones, Camera & Photo, Television & Video)	(Computers & Accessories)	0.609195	1.164167
3	(Headphones, Camera & Photo, Computers & Acces...	(Television & Video)	0.070903	1.114207
6	(Television & Video)	(Headphones, Camera & Photo, Computers & Acces...	0.022515	1.114207
25	(Headphones, Portable Audio & Video)	(Camera & Photo, Computers & Accessories, Acce...	0.013725	1.095370
20	(Camera & Photo, Computers & Accessories, Acce...	(Headphones, Portable Audio & Video)	0.108954	1.095370
11	(Headphones, Camera & Photo, Portable Audio & ...	(Accessories & Supplies)	0.350694	1.090892
34	(Accessories & Supplies)	(Headphones, Camera & Photo, Portable Audio & ...	0.004247	1.090892
5	(Camera & Photo)	(Headphones, Computers & Accessories, Televisi...	0.004804	1.027174
4	(Headphones, Computers & Accessories, Televisi...	(Camera & Photo)	0.306358	1.027174

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

# temp_db_cat = select_electronics_db.copy()
# temp_db_cat["concat_cat"] = temp_db_cat.category + "|"
# temp_db_cat = temp_db_cat.groupby(by="user_id", as_index=False).agg(
#     {"concat_cat": pd.Series.sum}
# )
# temp_db_cat.concat_cat = [i.split("|") for i in temp_db_cat.concat_cat]

# # to remove duplicates in the transaction, we use the 'set' type casting on the list before converting it back to list
# temp_db_cat.concat_cat = [list(set(i[: len(i) - 1])) for i in temp_db_cat.concat_cat]

# # this will have certain users where the user has purchases items from only one category and not from 2 or more categories
# # for forming association rules, we have to consider cases where user has purchased items from 2 or more categories
# temp_db_cat["cat_num"] = [len(i) for i in temp_db_cat.concat_cat]
# temp_db_cat = temp_db_cat[temp_db_cat.cat_num > 1]

# category_transaction_list = temp_db_cat.concat_cat.tolist()

# encode_cat = TransactionEncoder()
# array_transactions_cat = encode_cat.fit_transform(category_transaction_list)
# # items as columns can be seen via
# cat_cols = encode_cat.columns_

# # We form a dataframe of categories
# market_basket_cat_clean = pd.DataFrame(array_transactions_cat, columns=cat_cols)

# # We apply apriori algorithm to see the connection between items
# model_cat = apriori(
#     market_basket_cat_clean, min_support=0.0013, max_len=5, use_colnames=True
# )
# rule_cat = association_rules(model_cat, metric="lift", min_threshold=1)
# rule_cat = rule_cat.sort_values(by="lift", ascending=False)[
#     ["antecedents", "consequents", "confidence", "lift"]
# ]

# apriori_category = rule_cat.copy()

'Below code is to create the website for demonstration of Recommendation System, Hence it is commented out'

Modcloth Database

temp_db_cat = select_modcloth_db.copy()
temp_db_cat["concat_cat"] = temp_db_cat.category + "|"
temp_db_cat = temp_db_cat.groupby(by="user_id", as_index=False).agg(
    {"concat_cat": pd.Series.sum}
)
temp_db_cat.concat_cat = [i.split("|") for i in temp_db_cat.concat_cat]

# to remove duplicates in the transaction, we use the 'set' type casting on the list before converting it back to list
temp_db_cat.concat_cat = [list(set(i[: len(i) - 1])) for i in temp_db_cat.concat_cat]

# this will have certain users where the user has purchases items from only one category and not from 2 or more categories
# for forming association rules, we have to consider cases where user has purchased items from 2 or more categories
temp_db_cat["cat_num"] = [len(i) for i in temp_db_cat.concat_cat]
temp_db_cat = temp_db_cat[temp_db_cat.cat_num > 1]

category_transaction_list = temp_db_cat.concat_cat.tolist()

encode_cat = TransactionEncoder()
array_transactions_cat = encode_cat.fit_transform(category_transaction_list)
# items as columns can be seen via
cat_cols = encode_cat.columns_

# We form a dataframe of categories
market_basket_cat_clean = pd.DataFrame(array_transactions_cat, columns=cat_cols)

# We apply apriori algorithm to see the connection between items
model_cat = apriori(
    market_basket_cat_clean, min_support=0.09, max_len=5, use_colnames=True
)
rule_cat = association_rules(model_cat, metric="lift", min_threshold=1)
rule_cat.sort_values(by="lift", ascending=False)[
    ["antecedents", "consequents", "confidence", "lift"]
]

	antecedents	consequents	confidence	lift
3	(Dresses, Bottoms, Tops)	(Outerwear)	0.408913	1.430131
10	(Outerwear)	(Dresses, Bottoms, Tops)	0.340495	1.430131
6	(Outerwear, Tops)	(Dresses, Bottoms)	0.482918	1.229786
7	(Dresses, Bottoms)	(Outerwear, Tops)	0.247926	1.229786
1	(Outerwear, Dresses, Tops)	(Bottoms)	0.771956	1.203933
12	(Bottoms)	(Outerwear, Dresses, Tops)	0.151836	1.203933
4	(Outerwear, Dresses)	(Bottoms, Tops)	0.536135	1.171256
9	(Bottoms, Tops)	(Outerwear, Dresses)	0.212688	1.171256
5	(Outerwear, Bottoms)	(Dresses, Tops)	0.585666	1.157542
8	(Dresses, Tops)	(Outerwear, Bottoms)	0.192421	1.157542
2	(Outerwear, Bottoms, Tops)	(Dresses)	0.794229	1.132625
11	(Dresses)	(Outerwear, Bottoms, Tops)	0.138837	1.132625
13	(Tops)	(Outerwear, Dresses, Bottoms)	0.125480	1.118801
0	(Outerwear, Dresses, Bottoms)	(Tops)	0.868050	1.118801

"""Below code is to create the website for demonstration of Recommendation System, Hence it is commented out"""

# temp_db_cat = select_modcloth_db.copy()
# temp_db_cat["concat_cat"] = temp_db_cat.category + "|"
# temp_db_cat = temp_db_cat.groupby(by="user_id", as_index=False).agg(
#     {"concat_cat": pd.Series.sum}
# )
# temp_db_cat.concat_cat = [i.split("|") for i in temp_db_cat.concat_cat]

# # to remove duplicates in the transaction, we use the 'set' type casting on the list before converting it back to list
# temp_db_cat.concat_cat = [list(set(i[: len(i) - 1])) for i in temp_db_cat.concat_cat]

# # this will have certain users where the user has purchases items from only one category and not from 2 or more categories
# # for forming association rules, we have to consider cases where user has purchased items from 2 or more categories
# temp_db_cat["cat_num"] = [len(i) for i in temp_db_cat.concat_cat]
# temp_db_cat = temp_db_cat[temp_db_cat.cat_num > 1]

# category_transaction_list = temp_db_cat.concat_cat.tolist()

# encode_cat = TransactionEncoder()
# array_transactions_cat = encode_cat.fit_transform(category_transaction_list)
# # items as columns can be seen via
# cat_cols = encode_cat.columns_

# # We form a dataframe of categories
# market_basket_cat_clean = pd.DataFrame(array_transactions_cat, columns=cat_cols)

# # We apply apriori algorithm to see the connection between items
# model_cat = apriori(
#     market_basket_cat_clean, min_support=0.09, max_len=5, use_colnames=True
# )
# rule_cat = association_rules(model_cat, metric="lift", min_threshold=1)
# rule_cat = rule_cat.sort_values(by="lift", ascending=False)[
#     ["antecedents", "consequents", "confidence", "lift"]
# ]

# apriori_category = apriori_category.append(rule_cat)

# apriori_category.to_csv("apriori_category.csv")

'Below code is to create the website for demonstration of Recommendation System, Hence it is commented out'

7. Conclusion:

We have analysed of data, cleaned the data and build additional features from given data which helped us build our recommendation system.

Recommendation systems built in this notebook are based on the below parameters:

i. Ratings of the item & brand

ii. Best-selling items/brands

iii. Time-span of items/brand i.e. overall and latest trends

iv. Association Rules using Market Basket Analysis/Apriori algorithm for items & categories